Ask Paul: January 31 (Premium)

Happy Friday! January was a slog that somehow went by in a breeze, but I’m happy to put it behind me. This is a great way to start.

? History repeats itself

JaviAl asks:

In the early 2000s, when Microsoft introduced the .NET Framework, it announced the migration of all its products to this technology: Windows .NET, Office .NET, etc., which was brought down by the Longhorn disaster … After the reset, they put a ban on .NET in the development of the next version of Windows. They were prohibited from using .NET in all Windows development except Windows Media Player and games.

I literally just re-read this part of Steven Sinofsky’s book, Hardcore Software, and strongly recommend that you–and anyone else reading this–buy the book and read it as well. It’s quite insightful and, as important, is a first-hand account. But the short version of this history is that Microsoft went overboard in trying to compete with Java and the web as a platform by creating .NET a proprietary alternative to both and then proclaiming publicly that the entire company was going to pivot around this technology, down to the branding. Inside Office at the time, for example, they had no plans to integrate .NET in any meaningful way and were never going to “rewrite” the products in .NET, so they were as confused as anyone about that announcement.

But you ask about Windows. The reason .NET was dropped in Vista is that Microsoft only had 15–18 months to bring Vista to market after the reset, the Windows team was always firmly in the Win32 camp, they had to cobble together what they could from the mess of Longhorn off the floor, and they were confronted by a harsh reality: This thing was “bloated” from a resource usage perspective compared to XP and most existing PCs couldn’t run Aero Glass. So the decision was made to remove .NET from Vista because managed code ran more slowly at the time and this code would slow down Vista on low-end (common) PCs of the day that otherwise met the hardware requirements. (Microsoft still ended up facing a class action lawsuit over “Windows Vista Ready” PCs.)

There were numerous side problems triggered by these side-by-side efforts. Microsoft stopped developing or updating Internet Explorer because .NET’s Avalon technologies were going to replace HTML … or something. And this would all be part of Longhorn. But then it wasn’t, and Microsoft had to get that going again. It was a mess. But the goal was speed to market: Just get Vista out the door.

This decision had a chilling impact on developers who were adopting .NET because Microsoft had made it the center of its strategy and then quietly walked away from it in Vista. In the end, Vista wasn’t Windows .NET, it was just like XP, Windows with the .NET Framework tacked on the side. For whatever it’s worth, .NET did go on to “win” temporary against Java on the server, and the Server team at Microsoft embraced .NET wholeheartedly in ways the client team never did. The problems were all on the client side.

An Apple presentation Jobs took word for word the WinFS proposal with Spotlight, and Avalon with its new Aqua interface, and presented it as laughing at the disaster and delays of Longhorn. The difference was that it had more stable and functional parts than Longhorn. At Microsoft they were impressed with the WinFS cloning functionality in Spotlight that was already working. Apple seems to make life less complicated, tastes good, works better and gets better results.

Yeah, this history is rather incredible. A couple of points. I write about this in Windows Everywhere as well, of course, and the relevant bit is available in Programming Windows: Redmond, Start Your Photocopiers (Premium). But this speaks to a fundamental difference between Microsoft and Applet that I think persists to this day to some degree. Where Microsoft thinks in terms of “platforms,” these big, weighty constructs that require time and architecture and broad support, Apple thinks in terms of just getting it done. And if you go back and watch the OS X Tiger introduction at WWDC 2006 video today, it’s striking how they were able to just look at this problem pragmatically and say, look, we solved this in iTunes. Can we apply it to the file system? Meanwhile, Microsoft was trying to architect this heavy, database-based file system because that’s what good computer science demanded. Or something. (Bill Gates was obsessed with using databases everywhere.)

Not-so-random side-note. Coincidentally, I literally just watched a video, Oral History of Bertrand Serlet, that touches on this history as well. And he reveals that Apple had a similar issue with Objective-C and spent years trying to replace it while publicly promoting it as the answer to developers. (Serlet was an Apple executive who led the OS X team during that era and he hired/promoted Craig Federighi.)

Anyway.

This was a different era. In the years since, the smartphone has replaced the PC as the center of personal computing and we live in a far more heterogeneous world of online services, software, and hardware. Apple plays a much bigger role in personal technology than does Microsoft, which makes sense as that it’s focus. And Microsoft plays a much bigger role in corporate computing, the cloud, and now AI, which is an extension of the cloud, and that makes sense as well.

There are exceptions. Windows, which is a sort of legacy holdover, along with Microsoft 365, and Xbox, which is Microsoft’s only “pure” consumer brand. But for the most part, these companies, which once went head-to-head broadly, don’t really do so anymore. It’s interesting to me that the Mac, for all its success, never seriously threatened the PC from a usage or market share perspective. Nor have other desktop platforms like Chrome OS or Linux. Each certainly has its advantages.

Xbox is, of course, evolving into a software and services publisher, something more akin to Microsoft 365 than Windows. That makes sense in the context of Microsoft, but it was also forced on the company by market realities. I’m not sure it ever would have occurred to any in Xbox or elsewhere to go cross-platform or embrace services otherwise, not that it matters. But for all the hand-wringing about console sales, an Xbox that “meets customers where they are” just makes sense. Not just generally, but within Microsoft specifically. That’s the whole company. Except for Windows.

The Apple value proposition today is real, that’s why it’s so successful. This came up randomly today on First Ring Daily, because of the earnings announcement, but I happened to go all-in on Apple over the past year, and it was quite an “investment” (read: cost). But you buy-in for all the expected reasons and it mostly holds up. Nothing is perfect. And maybe I do understand why the Mac ever really threatened the PC because I still very much prefer Windows. But Apple pretty much nails it on consumer hardware and services. It’s difficult to argue with.

? Slice in time

JaviAl asks:

What do you think about the market share, according to StatCounter, of 62.7% for Windows 10 versus 34.12% for Windows 11 eight months after the end of Windows 10 support?

Heading into 2025, the big debate was whether Microsoft would extend support for Windows 10 past the announced October 2025 EOS (end of support) schedule. There’s so much that goes into this, but it’s interesting to me that Windows 10 in the end succumbed to the traditional 10-year support lifecycle that it was supposed to replace. (OK, only sort of. The version of Windows 10 that will exit support in October is 2022, so it will really only be 3 years old.)

Microsoft has extended support for two major Windows releases in the past, both with extenuating circumstances. Windows XP related to the question above, and the need to support netbooks with their low-end system requirements that made Vista a non-starter. And Windows 7, which Microsoft internally used as the response to anyone who hated Windows 8. But with Windows 7, Microsoft added three year of optional and expensive Extended Security Updates for businesses only.

But Windows 10 is unique because of its unusually high usage share at this time in the schedule. There are all kinds of reasons for that, we all get it. But most of the complaints one might have had about Windows 11 were resolved long ago. Microsoft is offering paid extended support for consumers and businesses for the first time. And by October 2026, one year from the EOS date, it’s difficult to defend still using a PC that is 9 years old or older. It’s time to move on.

And that’s what I think. It’s time to move on. Windows 10 is dated from a user experience perspective, and it’s starting to get dated technologically. With Windows 11 version 24H2, in particular, and especially with Copilot+ PC-class PCs with Windows Hello ESS, the technological gap is widening. What was once just marketing is now true. It’s time to move on.

And of the slight rise two months in a row, November and December 2024 of Windows 10 market share versus the drop in Windows 11 market share in that same period?

There’s no reason to think or worry about monthly changes in OS usage share. They’re not real, and this data is suspect to begin with. Windows 10 is not experiencing a usage uptick. A quarter is the minimum data range I think about, and the trends there are clear: Windows 10 usage is declining. Slowly, yes. But declining. It was in the high 60s, percentage-wise, a year ago, and in the low 60s today.

It’s likely that the unique nature of Windows 10 usage in early 2025 will just as uniquely disappear throughout the year, and rapidly. And we’ll see where it lands. But it doesn’t matter. Obsoleting the past is necessary, if only from a security perspective. It’s not reasonable to expect Microsoft to “just” keep supporting this thing. It’s literally from a different era, in the sense that the goal at the time was a cross-device apps platform that no longer exists, using a UI based on phone that Microsoft killed off a decade ago. It’s a weird truth of Windows and the PC that platforms that were not loved in the beginning, like XP and Windows 10, suddenly become these things people cling at EOL.

I look at the Copilot+ PCs specs, especially the security platform Microsoft just touted with the recent (and otherwise bland) Surface PCs refreshes, and see what PCs should be. It’s worth scanning the “Secure by design and by default” section in that announcement. This should be the baseline, and now it’s becoming that. Or, if you want to go all-in on the topic, this Microsoft post about the Pluton processor. (And the Windows 11 Security Guide and Windows Hello Enhanced Sign-in Security on Microsoft Learn.)

It’s time to move on.

? Nadella’s Folly?

train_wreck asks:

Could Deepseek upend the massive investments big tech has made in building out planet-melting datacenters for AI?

Short answer, yes.

But … will it?

There are still so many questions about DeepSeek, but what I keep coming back to is that it doesn’t really matter “how” they did it or at what cost, as the genie is out of the bottle. And to me, the most incredible admission in the recent Microsoft earnings was Satya Nadella stating that there will be new optimizations in Microsoft’s AI services “because of what DeepSeek has done.” Not just because of that, there are other optimizations that it and OpenAI are implementing too. But him calling out DeepSeek that way tells me it’s real. That these things will get better, and DeepSeek plays a role in that.

Two related points.

In Microsoft Earnings Analysis: FY25 Q2 (Premium), I point out something that Leo brought to my attention on Windows Weekly, the Jevons effect (paradox) which, in my words, “dictates that when technological resources become more efficient, overall demand paradoxically increases so much that total resources combined still continue to rise.” And Nadella referenced this effect, too, though not by name, noting that these efficiencies will simply lead to more AI consumption, which is good for a “hyperscaler” like Microsoft. In other words, whatever innovations DeepSeek and other optimizations do or do not deliver, this will not save the planet: We’re still going to need all those power-hungry datacenters because AI usage growth will outstrip the efficiencies.

Also, I retweeted Steven Sinofsky’s hot-take on DeepSeek the day that news hit, and he makes a good point. (Yes, I’m referencing Sinofsky again.) Which is that this kind of innovation always comes from outside the dominant players, because those companies–Microsoft, Google, and Amazon, basically–will never innovate when doing so threatens their business models. And the Big Tech cloud business model is to outspend everyone else and build infrastructures no one else can afford to build, limiting competition. In my simple brain, I view this as both “when you’re a hammer, every problem is a nail” and the Jurassic Park quote “life finds a way.” That is, unable to match the AI infrastructure build-out because of artificial reasons–in this case, US export restrictions–a company in China found a way. But this could have happened with a small startup anywhere, too, and still probably will. Life finds a way: Faced with limitations the dominant players don’t have, you’re forced to think differently.

Whether or how those two points collide is unclear. But I credit Satya Nadella for moving aggressively on AI, as it was a bold bet that will pay off, and it positions Microsoft for a new era of growth. I always have various concerns, in this case about whether the ubiquity of AI somehow undermines Microsoft as a business. But in the sense that Microsoft could always do well as backend cloud infrastructure, I feel like the same is true of AI. It will be OK. (It will be more than OK, but you know what I mean.)

? Office behaviors

spacecamel asks:

If we can go back to last week’s controversy and talk about office for a second. I was not happy with the price increase since Microsoft was taking away the choice of which AI that I could use by putting it into the price. There are a lot of plugins to allow for what I want. You had mentioned a few times that we should be able to go back to the “classic” version without the AI price increase. I looked through my Microsoft account page but do not see the option. Can you write a how-to with screen grabs to show how to select the classic option?

Microsoft documents this process on its Support website, but it is predictably–and, to be fair, understandably–non-obvious. You have to cancel your current subscription, and then select a classic plan when prompted. Stepping through this myself, I don’t see the cancel option, but I wonder if that’s because my Microsoft 365 Family subscription is extended out to mid-2028 at the moment.

Or .. one might argue this is some vague form of enshittification, or dark patterns, or whatever. For now, I’m just going to guess that it’s early days on this and that through time and/or customer feedback this will be made more obvious and/or start working. The real question is when the classic option goes away. But there, too, I don’t see it as malicious. Microsoft 365 Copilot is bolted onto Office today because of the way it was created, but it makes sense for it to just be integrated. That will happen in time.

? Photos, backups, and the cloud

Jollytiki asks:

I read your recent article about the time you had with Google. I went back through your “What I Use” and it stated you used Google Photos for photo backup to the cloud. I am sure you have gone through the trouble of downloading all of your photos from Google. I did it recently and it required me to use Google Takeout which for me was a disaster. The output from Google Takeout is small zip files with the pictures spanning across 20 or so files that have to be unzipped.

Yes.

So, about a year and a half ago, I underwent a massive digital decluttering effort that involved me consolidating all of my photo collections–from my NAS, OneDrive, and Google Photos, primarily–into a single collection that I then backed up to different places. I finished the photo consolidation one year ago, and I continue to back up phone-based photos to OneDrive, Google Photos, and, now Apple iCloud too, though the latter only occurs when I’m using an iPhone. For that reason, and because I have my wife’s photos to think about too, I will be doing annual clean-ups of the past year’s photos. That’s one of the tasks for this trip to Mexico. (Which I’ve only started, but I keep getting distracted by other work.)

Regardless of the details, it’s important to back up phone-based photos to at least two services, and my recent experience with YouTube/Google is one reason why. You can get locked out of an online account with no recourse and lose important data. It’s worth keeping a local copy as well, and that might involve plugging your phone into a PC and copying the photos wherever (a NAS, some PC at home with a removable drive, whatever). I don’t think it makes sense to not use online accounts, but I guess this is a personal decision.

What led me to using Google Takeout is this: After listening to Windows Weekly a while ago about what a disaster it would be to be locked out of online accounts and how a person can backup their cloud data to a NAS and have a mirror of all data on a person’s Google and OneDrive account to protect themselves. While setting up my NAS I realized there was no direct way to sync google photos with my NAS…they are not stored on Drive.

Yeah. This is a weirdism of Google Drive/Photos: There’s no file system view of your photo collection in Drive. That’s not the case with OneDrive: You can navigate right into your photo collection from File Explorer. I really prefer that capability.

All of the older photo’s from a few years ago showed up but at that cutoff date that they changed the storage location of photos. Therefore I had to go through the agony of using Google Takeout. After I realized this I moved all of my Google files and pictures to Dropbox. The app on my phone backs up all pictures to Dropbox with no problem and the price of 2TB on Dropbox is comparable to Google. Since I have not Asked Paul any questions yet. Have you run into this same problem or did I miss a step somewhere along the way and I was just torturing myself?

No, you are seeing it correctly.

I use Google Photos and OneDrive (and iCloud), as noted. But you can also mix and match with other services like Dropbox that make sense to you. I would just be careful to have at least two master copies of that data, for geographical redundancy, and for the possible online accounts issue.

I still haven’t pulled the trigger on a NAS, but I keep thinking about this, and researching the options. I would like a NAS that has a good phone app for photo backup, and I would ideally like to have two of these things, one in Mexico City and one in Pennsylvania, that are in sync, replicating that collection in both directions. This is unique to me, of course. But it’s the type of thing I’m always thinking about, regardless of last weekend’s drama. This data is too important to lose.

? Xbox hardware

christianwilson asks:

I agree with Microsoft’s strategy to bring Xbox games and services to more platforms. The writing is on the wall and it is the only reasonable path forward for this business. That said, there is going to be another Xbox console generation and the more I think about what that device(s) is, I can’t make sense of it. Who would buy a console without exclusives? No console platform has ever survived without exclusive games. If you could buy an Xbox that plays a bunch of games or a PlayStation that plays those same bunch of games and some Sony exclusives, the PlayStation is the right choice for almost anyone. Standout hardware features aren’t going to help because third parties largely won’t develop for it, especially if it’s not the lead console.

I just watched a couple of Phil Spencer interviews, and in one of them, he was asked about hardware, of course, and why anyone would buy an Xbox console. He answered that “on the hardware side, I want us to innovate to make the hardware the differentiator, that we got into this space where the differentiation on the hardware has gone down [compared to PlayStation] and it’s really been locked games that you’ve tried to make the identity of the hardware.” That is, of course, a reference to the Sony strategy. But from there, he launches into handhelds again–he is really telegraphing a coming Xbox/Windows handheld gaming device these days–and “the unique things that hardware manufacturers do, and I want our hardware to compete on power, and on innovation. So let’s have our platform [meaning Xbox everywhere] continue to innovate with service, and hardware work we’re doing, whether its control, whether it’s power, mobility, whatever it is that we’re focusing on. And let’s have our game reach you wherever you want to go play. That’s our identity going forward.”

That whole interview is worth watching, by the way. But my takeaway is that the next Xbox will clearly be mobile. And as we’ve discussed in the past, Microsoft is kind of stuck right now, in a holding pattern, or whatever, waiting for several pieces to come together. My belief is that this handheld Xbox will be Arm-based, for starters, and with Qualcomm or Nvidia graphics, but the details don’t matter. (AMD is allegedly working on Arm, too, and it does power the current-gen Xbox and PS devices.)

Phil Spencer has talked up an Xbox handheld multiple times despite Microsoft’s OEM partners already making Windows-powered handhelds. A Windows PC that functions like a Steam Machine isn’t hard to imagine and given PC gaming’s increasing market share, I think it’s just a matter of time before that happens. An Xbox that is a Microsoft-designed gaming PC would fit in well in that market.

100 percent. But the question is how they get there. There will be Windows options, obviously, that’s happening, and the possibility of Xbox doing work with Windows to improve that. So maybe there are Windows and Xbox portable gaming devices. Each makes sense in its own way. It would be wonderful to gain access to Xbox games on Windows. Could we get Windows games on Xbox? Do these platforms “merge”? This is all speculation for now. Xbox is or has been waiting on something. And we are waiting on them. Spencer’s continuous hints and comments make this clear.

Is Xbox, the console, destined to become the Surface of the gaming PC space?

I guess that depends on how you view Surface and its role/position in the PC space. In consoles, Xbox is a distant number three, but it’s not as much of an also-ran as is Surface. Surface PCs–and Xbox consoles–are good hardware, well-made and with modern designs. But they don’t sell well. Xbox is heading in this new direction, and that’s not something Surface can do. Microsoft either continues making Surface PCs or it doesn’t, and as much as I love Surface, I feel like it’s just never clicked and maybe doesn’t make sense as a business. But Xbox does make sense as a business, it’s just that that business is a software and services provider that, for now, still makes hardware too. Phil Spencer and Xbox keep telling us important Xbox hardware is, but every Microsoft earnings report tells us how unimportant that is, too. So they’re trying to thread that needle in appeasing the hardware fans while continuing to remind us all that this is about meeting us where we are. And increasingly, Xbox gamers are not on that one console.

I could see Microsoft exiting the PC and console hardware businesses. There’s a good argument to be made that both are a drag on earnings and that neither can ever be profitable. Hardware is low-margin (unless you’re Apple) and has all kinds of other problems. Microsoft’s strengths are in software and services at scale, not hardware. And so on. These feel inevitable. There are fans who wish otherwise. Including me. But logically, pragmatically, the end game is obvious. Perhaps there is a future where hardware makers sell Xbox-branded devices of whatever kind. Maybe that needs the Windows/Xbox “merger” to happen. If it’s even real. All we can do is speculate. Until we finally find out what’s really happening.

? Two lines meet on a graph

helix2301 asks:

I read the premium post on earnings. My question is: Microsoft saw a dip in gaming profits. As they sign more people up for Game Pass, will they continue to see this? I notice myself not buying as many games because of Game Pass. Will putting xbix games everywhere offset this? Has Microsoft ever mentioned the cost of Xbox cloud gaming cause people streaming games is an expense?

I was surprised by overall gaming performance in the earnings report, especially in the wake of that report about Microsoft being the biggest video game publisher in the world now. But even without that, it was reasonable to expect another drop in hardware revenues, another gain in services tied to Game Pass adoption, and, what I expected, a gain in gaming overall.

But what is that based on?

I didn’t cover this in that earnings post because it kind of took on an AI-based life of its own. But it’s important to understand how Activision Blizzard and the shift to Game Pass services impacts the business over time. And it is important to Microsoft that it not give us that information unless it is exceedingly positive. This is the classic magician’s tactic of waving the hands to distract the audience, but applied to financial reporting.

In previous quarters, Microsoft provided a slide in its earnings presentation called “Net impact from the Activision acquisition.” But this was the first quarter that it did not do so. Here’s the version from the prior quarter.

In short, the net impact for now is negative: Microsoft essentially paid cash for Activision Blizzard, but there are ongoing costs, some of which are merger related and temporary, and some of which are ongoing, the normal cost of doing business. Over time, as redundancies are reduced and the temporary acquisition costs go down and then disappear, the goal is for this business to be profitable, which it was as a standalone business. In July 2023, I looked at how Activision would impact Xbox/Microsoft Gaming and Microsoft overall from a financial perspective if the two were a single company. And not including the costs of this transition, that’s pretty much what we’ve seen so far. A big deal for More Personal Computing and Gaming. And a relatively minor bump for Microsoft overall. But a net loss overall. Because costs and expenses.

With this quarter, we can no longer see how the acquisition is impacting the business and the year-over-year comparables are with the Gaming segment that already included Activision Blizzard. The 2 percent growth we got in Xbox content and services was attributed to Game Pass, but it’s really because of Activision: This was the first quarter in which gamers could get the new Call of Duty and Game Pass and many clearly did the math and decided that was the way to go.

The question, generally, but also what you’re asking, is when these changes intersect into a net positive for Xbox/Microsoft Gaming. That is, at what point does Game Pass become popular enough that it offsets whatever losses to individual game purchases that the subscription triggers? I guess we could look at this like the shift from physical game media to digital. But you have to wonder whether we’ll see something similar to the situation in the music industry, where physical media was far more lucrative. That market shrunk as it moved to digital and then subscriptions, but the point I was trying to make in the recent Spotify story is that it was doing to basically disappear otherwise. Music as a business is not what it once was.

But gaming is growing. And so the hope here is that this does make sense. And it almost certainly will. Shifting from a hardware focus to a services focus makes sense, generally speaking. Going cross-platform broadly absolutely make sense. And at some point the entire business makes sense. We’ll hear when and if that happens. In the meantime, Microsoft will hide the data and cherry-pick from the good news each quarter. The Activision net impact transparency is over.

Everyone’s habits are different. In the past, I wasn’t buying too many games outright. I was playing Call of Duty. Pushing COD onto Game Pass makes that service interesting to me, and it opens me up to more games, a lot more. And so maybe my usage grows and expands to new games. Instead of spending $60 or $70 each year on one game, I spend $20 a month. That’s a big win for Microsoft. And it’s OK for me if I’m using it. If I’m not, I guess I could just buy the games I want. But Game Pass is good exposure to other games, as is Xbox Cloud Gaming. When these things work everywhere, or at least in more places, all the better.

Going all-digital always made sense to me. Going all-subscription is less clear, and I don’t see that happening, and certainly not anytime soon. We’ll have a choice of how to acquire games. But having a choice of where to play games is great too. I think this whole business model makes sense. But again, it’s a matter of time, and we don’t get to see the progress unless there’s good news to report. We see bad news at a high level. And we never get to see individual costs.

? Screen recording

jrzoomer asks:

Paul I read over your recent What I Use (January 2025) and wanted to know what you use for screen recording in Windows. Do you use the built in screen record function of the Windows Snipping Tool? Or do you use the NVIDIA overlay (or XBox Gamebar)? Or do you use dedicated software programs like OBS, Snagit, ShareX etc?

I use OBS Studio for screen recording, at first because this was what TWiT required of me for recording Hands-On Windows. But I’ve come to really like it. OBS is free, it’s available in the Store, it works fine on Windows 11 on Arm, and now that I’ve used it a bunch, I can configure it correctly on a new PC really quickly. That said, Snipping Tool does work well and would likely meet most needs.

I don’t do a lot of recording in games, but I have used Game Bar for that, yes, and I use Game Bar to pin the Performance widget with just FPS showing so I can see how COD performs on different PCs.

I use Greenshot for screenshots. I’ve always loved it.