Ask Paul: July 10

Happy Friday! I was expecting Xbox existential dread this week and was happily surprised it didn’t happen. But here’s a great set of reader questions to keep our minds off those troubles and get the weekend started a little early.

“Layoffs begin at Ubisoft following Assassin’s Creed Black Flag Resynced launch”

Thanks for helping us get that out the door, guys. Now show yourself the door

🔮 When will then be now?

NickTech asks:

Paul, what are your thoughts on the news that Microsoft is developing its own AI LLMs? Do you think Microsoft can put out something that is just as good (or even better) than the current LLMs in use, or could this cause people to migrate off Copilot?

Microsoft has been working on its own LLMs for years, but the drive to create in-house “frontier” models (as we now call them) that can compete with the best took off when Microsoft and OpenAI reconfigured their partnership earlier this year, giving them (Microsoft) the ability to pursue super-intelligence without OpenAI’s help or involvement.

The news this past week is that Microsoft has been transitioning to its own models more than we previously knew, though as I wrote that story, I realized that it was less dramatic than the headline of the original report suggested. Microsoft has been open about this transition since Build 2026 in May, when it announced several new top-tier Microsoft AI (MAI) models.

Building AI models that can compete with the best in the market can’t happen overnight, even for Microsoft, which has been using and enhancing OpenAI’s best models for over three years. In some ways, this situation is a bit like the component crisis, where even the most well-equipped silicon makers can’t just spin up a new memory fabrication facility and solve the problem quickly. But this is software, and Microsoft does have the expertise and, as crucially, the infrastructure to pull it off. So I feel like it’s inevitable that Microsoft is competitive in this way.

But competitive is somewhat subjective and varies almost by day because new model releases always reset the rankings across whatever various workloads these models are measured against (coding, reasoning, generation, etc.). Gemini may or may not be competitive today compared to the latest and best models from Anthropic and OpenAI on any given day or week, but the overarching issue is just being competitive because the usage of Google AI will be driven largely by its customer base. Many/most who use Google Docs and Workspace, or whatever, will stick with Gemini because it’s easier and more seamless, works everywhere they work, and is a single bill.

Microsoft can and will benefit from this same dynamic with Microsoft 365. You can call it inertia, but whatever the name, Microsoft is dominant with the enterprise, and as long as Copilot isn’t a complete laughingstock, some percentage of that user base will stick with it. Granted, that’s been a small number so far. But there are dynamic shifts happening that will benefit Microsoft. On-device and smaller cloud-based models are getting much better and are reaching a “good enough” area. Their models are getting less expensive to run (for Microsoft) and will remain so as the industry shifts to usage-based pricing. For many customers, the free models will just work, keeping them in Microsoft’s ecosystem. And so on.

The timing is unclear.

I made the point previously that Microsoft AI CEO Mustafa Suleyman is the only human being I trust in Big Tech AI, and that still holds; the rest of these jokers are like villains from a Batman comic. And his push for super-intelligence over AGI is, I think, more realistic and more morally compatible with giving a crap about people and not just growth and profits. But this is still a relatively new business. Microsoft AI shipped its first models less than a year ago, and it took a year and a half for that to happen.

The problem for Microsoft isn’t getting there; it will. It’s that it may get there after it no longer matters. For now, Microsoft–and Google, too, to a lesser degree–have that built-in inertia of a user base that expects to use traditional productivity apps and services (Word, Excel, etc.) to get work done. Big Tech and Big AI companies are building ways for their models to interact with these apps and services, and many other things, and the methods range from unsophisticated (basically screen scraping, which is all computer use really is) to deep integration. On the flip side of this transition, Microsoft and other app makers are (or are not) extending these legacy productivity tools with more seamless ways to interact with AI. They are exposing features from within apps that AI can manipulate directly.

The end game, I think, is that those legacy apps and services largely disappear. That won’t happen overnight. And they will live on as back-end services for AI, whatever that looks like. But the traditional workday, where you sit down at a computer, have some task to perform, and think about what app or apps to use to do that, starts to disappear. Increasingly, the AI is the user interface, and we will use natural language to describe the task at hand, and the AI will orchestrate the use of whatever tools to get to the same result. Which may be a document of some kind, or not. But it will do so without running whatever apps, and you will no longer even need them.

In that world of the future, Microsoft’s inertia advantages start to disappear. Having Microsoft 365 is no longer necessary and no longer the source of lock-in for customers. They can just pay for AI and use that for those tasks and for numerous other things, too. Apps like Word and Excel will be relics of the past, but all the features they contain will still be out there somewhere, powering AI models.

So, yeah. Timing. That future is already unfolding. The two-tiered news yesterday about OpenAI combining ChatGPT, Codex, Work, web browser capabilities, and everything else it does into one super-app (ChatGPT) and killing its standalone browser tells us two things. OpenAI is making it up as it goes, which its competitors all do too, and we already knew. And as time goes by, there is the increasing understanding of where users will spend their time. The AI assistant becomes the user interface. That’s the OpenAI super-app vision. As I wrote last October, AI is the end of apps.

To be fair to Microsoft, it, too, saw this future, and it may have been onto this earlier than OpenAI. You will recall that Microsoft CEO Satya Nadella once described Copilot as the new Start menu, and how we all lost our minds. But you may not remember that he said that almost three years ago. And Microsoft is also working on its own super app, which it confirmed during the Build 2026 keynote when it introduced its Scout personal AI agent. I suspect this will be similar to the evolution of ChatGPT.

Long story short, Microsoft has to be competitive in frontier AI models; it has the infrastructure to do that effectively for itself and customers, and cost savings are just one solid outcome. The tricky bit is that this work needs to make up for losing Microsoft 365, essentially, which is currently its single biggest business.

If you think about the cloud era and how that defined Microsoft for over a decade, Windows was diminished, but Microsoft 365 benefitted from that push and grew. But if AI plays out the way everyone, Microsoft included, believes, then Microsoft 365 disappears. So this effort needs to generate revenues commensurate with Microsoft 365 for that to be OK. It may just be that Microsoft 365 evolves to this place where this Copilot super app, whatever it’s called, is the center of things, just as Outlook or Teams might be today, and legacy apps stick around and are used by a dwindling user base until they just die out.

Or it may not. AI could be an existential threat to Microsoft as a business. Which explains why it invested in OpenAI, pushed AI so aggressively, and is now building its own frontier models. Microsoft sees this, literally, as do or die. I disagree with some of the things it’s done over the past three years or so, but I do think that assessment is correct.

“iPhone 18 Pro and 18 Pro Max are rumored to get bigger batteries this year”

How big were the iPhone 18 Pro batteries last year?

🪟 Windows is objectively better

jrzoomer asks:

Paul what keeps you on PC as your primary platform as opposed to Mac? Given that most of everything we do these days is in the web, and most popular locally run applications run on both platforms.

The knee-jerk reaction most would have to this question is to just say it’s familiar, that 30+ years of usage and evolving workflows along with the platform create that inertia I noted in the previous answer. And that probably does explain why some stick with Windows. For others, it may just be that they’ve moved on so that they are using Windows less frequently and just don’t worry about this kind of thing. Why go to the trouble of switching (to a Mac, Chromebook, Linux, or whatever) when you only occasionally need the big screen and keyboard and the PC you already have works fine for that?

Looking at me from the outside, you could also easily claim that I write about Microsoft and Windows primarily, that this is my career, and that not using Windows would be some form of career suicide or whatever. This I also understand, though I hope that anyone who reads what I write understands that I’ve never approached this as a tunnel-visioned cheerleader, am mostly interested in helping other people with technology, and routinely use other platforms, apps, and services in ways that most do not.

None of that is why I stick with Windows. I stick with Windows specifically because I routinely test other platforms. Windows, for all its faults, is still the best desktop personal computing platform, and I mean that objectively. Not because of familiarity or some subjective preferences. It’s just the best there is, and it has withstood every competitive threat I’ve thrown at it. Literally, over 30+ years.

This is a combination of things. It is the competition not beating Windows in key areas despite improving dramatically pretty much across the board, and more so recently than ever before. And it is despite Microsoft undermining the experience in Windows through enshittification. No one is more aware of both of these things than I am, maybe literally. But Windows is still better.

When I review or test anything, there are some universal comparisons that matter. When you switch platforms, say between an iPhone and a Google Pixel phone, as I do regularly, there are things that are better on either. Things you miss from the other. Things you appreciate on the device you’re currently using. Etc. The choice one makes–most people would just pick one or the other and just stop thinking about it–is based on some weighted matrix of factors, personal to the individual, some score you arrive at where one wins and one does not.

I never run into a blocker in Windows. I always run into blockers on other platforms. Literally always. There are workarounds and fixes. There are platform improvements that shift things upwards for the competition to some degree and close the gap. And I keep waiting for some platform to just nail it and surpass Windows in some way. And it just doesn’t happen. Not yet. Maybe it will, and I am certainly open to that. Am, in fact, actively seeking it out because that is what I do.

Is it the software or hardware primarily?

It’s both. This is subjective and specific to whatever work we do and how we work, but one of the key strengths of the PC industry, obviously, is the choice we get in not just hardware makers but also all the different models, form factors, and whatever else. If I want the biggest screen possible on a laptop, I have an incredible range of choices. But if I stick with Apple, I get a MacBook Pro 16-inch that is incredibly thick, heavy, and expensive, or I can compromise with a smaller 15-inch MacBook Air. Etc. There is no Face ID on Macs, and the only iPad that supports this is the most expensive model. It goes on and on.

If it’s software, after you try a Mac what things do you “miss” about Windows?

It’s always the little things, and to be fair to Apple, the Mac comes closest to Windows for the sweeping range of capabilities and workflow compatibility that I prefer. I can install OneDrive, Google Drive, or Synology Drive, and they work as well on the Mac as they do in Windows. All the key apps I want to use are there. Etc. But where the Mac falls flat from a day-to-day usage perspective, to me, is across a wide range of little things.

For example. There’s no real full keyboard navigation in all apps, like we get in Windows, and that’s exacerbated by third-party apps and how inconsistent they are. I use Affinity for photo editing, and this new version of the app has some changes I do not like, one of them being that it changed the keyboard shortcut for exporting, which is always the end-game for me. There was never a way to do this with a keyboard shortcut on the Mac, but on Windows, I can configure Ctrl + E for export. On the Mac, I cannot. (Not within the app.) But that’s just one app, and the keyboard navigation/shortcut thing is a system-wide problem.

There’s more, some of which you can fix with third-party utilities, as I do, like with Cmd + Tab not displaying all open windows and how maddening it is that each app has a different way to navigate between multiple windows. That keyboard shortcut to move between full-screen, normal, and hidden display modes is inconsistent between apps, and it’s maddening. The nonsense with full-screen apps each being a virtual desktop and how just using that means you can’t multitask normally with floating apps on the desktop. And then there’s Stage Manager, this bizarre thing tacked on the side. It goes on and on.

The way this goes for me on any platform is that I start using it, I install apps and updates, configure access to my data using whatever methods, and can deal with some limitations. But then I run into a little interruption in my workflow, like being forced to use the mouse to export in Affinity. And things like that keep happening. And it’s frustrating because some cannot be fixed or worked around; they are just limitations. And they add up. And I go back to Windows, where these limitations do not exist.

The problems with Windows are real. But after decades of using this platform and helping others workaround or fix its problems, after doing the work to document fixing the latest issues, the enshittification, leading up to De-Enshittify Windows 11, nothing has changed. The problems with Windows can all be fixed, and it’s not time-consuming or difficult. The benefits of Windows remain. The competition, always improving, still falls short. Maybe that changes. I’m ready to switch when it does. Happy to stay where I am if it does not. I just need to get work done.

I should mention this is a highly technical audience. Someone will look at some specific example above and know about some way to workaround or fix it, and that will be evidence to them that I am wrong about everything. But I only gave a few examples; the effort one would go to even try and fix all the problems with any of these alternatives is too tedious and time-consuming. And objectively, again, not all of them can be fixed. This isn’t something I made up 10 years ago and stopped thinking about. This is something I’m testing and probing all the time. Anyone who reads this site knows how I perk up whenever there’s some platform improvement out there. I am paying attention.

To each his own, of course. But that’s where I’m at.

Or if hardware, do you think the gap is closed now that Snapdragon seems to be catching PC up to Mac M series line of CPUs, or widening do to hardware costs of components?

Snapdragon X absolutely narrows the gap, but this is an interesting area because it depends on what matters to you. Everything is a compromise. Apple Silicon is magical for all the obvious reasons, but the trade-off to unified RAM is multifold; you can’t upgrade it after the fact, and there are no dedicated graphics. So it will be behind the PC in some ways, while being ahead in others.

What Snapdragon X accomplished was to give PC users, who are accustomed to choice, a new choice that provides most of the benefits that Mac users get with Apple Silicon. Better efficiency, battery life, and reliability, and incredible performance for day-to-day work. But neither is great for gaming or high-end scientific, engineering, and whatever other workstation-type workloads. We have  choices. Apple has more limited choices, and they do not address all needs.

In time, one hopes that Arm is a better supported target for games and whatever else on both PC and Mac. But for now, Snapdragon X is as close as we can get to Apple Silicon. It’s not as good as Apple Silicon in almost any area, but it’s good enough. It’s in the ballpark. And it’s dramatically better than x86, which needs to die by fire as quickly as possible but won’t because of inertia, misinformation, and habit.

I prefer Windows over other platforms. I prefer Windows 11 on Arm on Snapdragon X above all else. This is pretty much true for anyone who actually uses that type of PC. It’s eye-opening. We get so used to the nonsense in x86 computers that we just think it’s normal. But it can be better. And if you get a Snapdragon X/X2-based PC, it is better. To me, it’s the best.

“Scientists: No, we’re not living in a simulation”

Duh and/or hello

💻 What’s old is new again

helix2301 asks:

I was wondering if you heard all the buzz this week about the Microsoft Global Device ID. This is an ID on every windows pc that is used for cloud infrastructure. Now, security people are going crazy over it, very similar to the recall craziness a year or 2 ago. Wondering about your thoughts.

My initial reaction was twofold. First, that this isn’t something new. And second, that this is an example of fingerprinting, a technique so many Big Tech and other companies use to track users even when they’ve done everything they can to prevent that. The Cover Your Tracks measures this for web browsers–good luck finding an extension configuration that results in a truly randomized fingerprint–but this is a big problem on mobile, too. You can opt out of tracking globally on an iPhone, but Meta uses so many bits of data to fingerprint you that you may as well just give it your data directly. (I’m exaggerating; do not do that.)

Windows 10 and 11 don’t let you turn off telemetry entirely. There are third-party utilities like Win11Debloat that do let you do that. But the Microsoft Global Device ID is not part of that. We don’t know some aspects to this story, like whether the hacker had at one point signed with a Microsoft account, forever tying that combination of Windows software install and hardware device to him/that account. But I suspect that doesn’t matter, that it’s just the GDID somehow being transmitted whenever anyone using that PC accesses whatever online services. It’s not clear.

The good news, such as it is, is that this has been publicized, and now hackers can work around this form of tracking. Perhaps tools like Win11Debloat will be updated to block this too, or maybe we need new utilities.

This is a weird gray area. Few would complain that a hacker who took part in cyberattacks on airlines was caught. But it’s reasonable to ask about the broader implications of how he was caught. I find myself more outraged that Apple won’t help the government unlock a terrorist’s iPhone than Microsoft did help the government capture a hacker. But this is the type of thing that’s dogged Microsoft since Windows became popular, this belief that it contains backdoors for governments. Which, while easily countered by evidence (including the source code availability for all governments), is still a concern. As it should be for any Big Tech company.

The intent here is unclear. It’s unlikely that Microsoft intended to use the GDID for “surveillance” in some nefarious way, but rather as part of solid data gathering for telemetry of all kinds, mostly related to product reliability. But the worry that governments or malicious hackers could use this for literal surveillance purposes is reasonable. As is Microsoft cooperating with law enforcement in this way.

This sounds naive, I know, but this kind of thing doesn’t concern me personally. I’m not worried about the FBI or whatever knocking on my door or about getting pulled aside when traveling internationally. That kid–and my god, he is just a kid–should have been worried about those things because he was breaking the law explicitly and on purpose. I watch true crime shows all the time, and it’s shocking how many of the bad people in these stories didn’t understand how their activities could be tracked so easily using cell phone data or whatever. This feels very much like that.

I wonder if Dave Plummer will tackle this topic. Or Raymond Chen.

“How Cracker Barrel saved itself”

I assume it was the biscuits and gravy

⌨️ Simply the best

jrzoomer asks:

Paul are you still using a Microsoft Sculpt Ergonomic keyboard and mouse? If I remember correctly that was your go to

Yes. This is by far my favorite keyboard/mouse combo, and we have at least one set here in Pennsylvania and in Mexico City. There are all kinds of similar setups, and most are at least pretty good. But I’ve found one I prefer to the Sculpt keyboard and mouse.

“Microsoft fixes an annoying issue with Copilot in Outlook”

That it exists?

📖 It’s all about balance

OldITPro2000 asks:

Hey Paul, I was reading your Siri article earlier and didn’t realize I was logged out. I was able to read at least 90% of it before hitting the beginning of the text fade out and nearly 99% of it before I saw the Premium notice. I tried on multiple browsers and with ad blockers on and off with the same results. The fade out used to occur much, much earlier.

I don’t remember when we did this, but a year or two ago we increased the amount of text anyone can read in Premium articles.

It seems like this is either a bug or was intentional. Is this something you needed to do so articles are indexed more completely in search engines and chatbots?

It was intentional. Premium is a double-edged sword in the sense that you want to provide whatever value to those who pay for the service, but you also have other needs, one of which is visibility on Google Search and other indexing services. I don’t think about this kind of thing pretty much ever, but Robert came to me at one point and told me that Google had changed whatever ranking algorithm and that having so many walled-off Premium posts was now a problem. So we talked it over for a bit and collectively decided to increase the visibility outside of the paywall to meet Google’s needs. With the added theory that most Premium posts are quite long. Not all are, of course, but that’s OK. You can rationalize this kind of thing in different ways, one being that maybe some people will read what they can for free and then one day decide they would like to pay for it.

I’m not good at marketing or whatever. But that’s why.

“Microsoft Exchange Server on prem gets a little harder to use”

Yeah, it was always so user-friendly. Cough