Ask Paul: January 2 ⭐

Happy Friday, and Happy New Year! We head into 2026 full of uncertainty but also full of hope. So let’s kick of this holiday weekend, and the New Year, on a good note.

? Bad memory, bad

MartinusV2 asks:

While there are high memory prices right now, will that impact the next release of Qualcomm’s next Snapdragon x2 success?

I’m not sure that it will impact Snapdragon X2 more or less than it will whatever Intel and AMD are releasing, but I suppose that depends on a variety of factors.

Snapdragon X was widely adopted by PC makers in the sense that all the major players had at least a few models, with more arriving over time. But availability is also a fairly limited because it’s essentially a new and unproven platform and Snapdragon X probably accounted for less than 10 percent of PCs that shipped last year. So that might factor in for the new generation, and we’re just days away from CES 2026 and what rumors tell us will be the first Snapdragon X2-based laptops.

If these appear in market sooner rather than later, I have to assume that the orders were made long ago, so RAM pricing might not factor in all that much. But longer out, that could be an issue for sure. But also an issue for all chips and PCs.

Forced to bet, I don’t see the RAM pricing problems impacting Snapdragon X2 disproportionately. These will mostly be premium PCs with better margins, and it feels more likely that the biggest issues will be with more affordable PCs. I could be completely wrong on that.

Tied to this, I am curious whether we’ll see upgradeable RAM on more laptops, especially. Despite all the right to repair gains, RAM remains the one basic component that’s typically non-upgradeable on modern laptops. And this would be a good place to save some money. A customer who wanted, say, 32 GB of RAM could order a laptop with 16 GB and then upgrade later when the prices are better or enough time has gone by. Regardless of the current crisis, I was curious we haven’t already seen more of this.

With the memory problem, will we see a new revolution in optimization in software that will be more efficient memory or better code optimization?

If it continues for years and/or is permanent, yes. This is one of those “life will find a way” things, similar to what DeepSeek is doing with cloud AI. If you have infinite resources, you’ll use them, but people get inventive when there are restrictions or limitations.

Oddly, this is perhaps one area where AI could help: We could use GitHub Copilot and other AI pair programming tools to refactor code specifically to be more efficient. We should be doing this anyway, of course.

✔️ Mi amiga

wright_is asks:

Thinking of what-ifs… Just reading your year in review and saw the Amiga Forever article … I always wonder, if Commodore could have sorted itself out and actually marketed the Amiga properly, could it have succeeded, been the 3rd operating system, or even have supplanted the Mac? It had the potential early on, but was just hung out to dry.

Anyone who owned and most people saw any Amiga must wonder the same. I couldn’t have been a bigger fan of the platform, and I also had an Amiga 500, which I tricked out with a 1.3/2.x ROM toggle switch and a PCMCIA hard drive. I always lusted after the more impressive Amigas, like the 2000, 3000, and then 1200, but I could never afford them. And the Amiga 500 was truly impressive regardless, it a multitasking, game-playing marvel of technology so advanced in some ways that it seemed like something given to us by aliens.

The Amiga wasn’t perfect. The chipsets were originally designed for standalone videogame machines and there was a confusing array of often very limited graphics modes. This heritage made the Amiga well-suited for video production in an era when most computers had text-based interfaces, exaggerating its advantages because this was a non-mainstream use case. By the time Windows 3.x and VGA screens had arrived, I was jealous of how clean and high resolution those displays could be.

But the Amiga’s biggest problem, as you allude to, was Commodore. If you’ve not read Brian Bagnall’s excellent series of books about the history of this company, I strongly recommend all of them. It’s most depressing, of course: Commodore rescued Amiga from Atari (and former Commodore CEO Jack Tramiel) and then squandered the technology and products for the next several years, thanks to the idiotic non-leadership of Irving Gould. It had so many chances at redemption, but he couldn’t stop tanking the company.

If we can ignore that, and we have to, because otherwise this what-if makes no sense, then yes, I can see a timeline in which the Amiga kills a stillborn Macintosh because it has color and dramatically better technology, and the creative market that Apple fostered in the wake of Aldus Pagemaker and the laser printer belongs to the Amiga instead. And because Windows and the PC hadn’t happened yet, it could have taken over there as well, because the Amiga was price sensitive enough to negate the need for such things. And Bill Gates, instead of recommending to Apple that it license Mac OS, might have turned to the Amiga instead. We’d be free of x86 and free of the Mac’s high prices. And we might have avoided the Microsoft dominance and antitrust issues of the 1990s because it would have focused on applications. It’s wonderful and terrible to consider.

There is nothing sadder in our industry than this missed opportunity. But again, the blame must lie with Irving Gould. Commodore was too dysfunctional to ever have that level of success in the mainstream personal computing market that exploded in the 1990s. There is also some interesting alternative futures involving some number of what we used to call home computer platforms, like the Commodore 64/128, Apple II series, Atari 400/800/XE, and so on, improving in a manner similar to the IBM and DOS/Windows over time. I often wonder what Apple would have been like had it improved the Apple II instead of starting over with the Mac, and though it was in some ways underpowered, the Apple IIGS is a small glimpse at the possibilities.

This is all so depressing.

Would we be talking of Jay Miner in the same hushed tones as Dave Cutler?

Yes. And if you know anything about him and his work, you probably do anyway. Miner was the Steven Wozniak of Amiga, but more advanced, really. He implemented multi-chip computer systems with the Atari 8-bit line and then the Amiga, and real multitasking with the latter, and he did all that before 1985 when the PC and Mac were jokes, technologically speaking. His passing was a tough loss, though there were people at Commodore later with the skill to advance the chipsets who were thwarted, as always, by Irving Gould.

Programming the Amiga in Lattice C++ was also a dream, at the time, compared to Windows or Mac.

Exactly. At this point in time, Apple’s programming platforms were all Pascal-based, and Visual Basic, Turbo Pascal, and Turbo C hadn’t happened yet on the PC side, and Microsoft’s tools of that era were primitive. The Amiga also had a terrific hypertext system called AmigaGuide, a Rexx interpreter called ARexx, and even OOP/C++ class libraries by the 2.x era called BOOPSI. All this stuff was so under the radar and yet so incredible.

? Highs and lows

will asks:

Hi Paul, I recently listened to Steve Ballmer’s Acquired interview from last year, and I was struck by how candid he was. He essentially admitted he took his eye off the Windows desktop ball, called Longhorn his biggest mistake, and pointed to Azure as one of his biggest successes. It was a great long form interview anyone should listen to.

Yep. I am a big fan of Steve Ballmer and that interview pretty much explains why in a nutshell. No one is perfect, but his head and heart were always in the right place. He’s unfairly blamed for some things that were not his fault, and doesn’t get enough credit for those things he is responsible for (like Azure).

With sentiment around Windows today, even among longtime fans, seeming to sour lately, do you think Microsoft will make any meaningful Windows course corrections in 2026? If so, what’s the most likely area: the Windows experience, servicing and release cadence, or how it’s positioned alongside Copilot and AI?

I feel like we need to put the issues with Windows in perspective. The most controversial issues are just just enthusiasts pining for the old days, when everything was simpler, Windows was the center of the universe, and every announcement was interesting. But even many of the issues I complain about, the enshittification problems, probably don’t bother most mainstream users. As a community, we are getting smaller and we are digging in harder, and I think we sometimes lose sight of the big picture. If you really are bothered by whatever issues in Windows 11, there are fixes and workarounds for just about all of it. So I try to focus on that.

That said, I would love to see Microsoft make Windows great again, so to speak. I have been asking and begging for an official way to de-enshittify Windows for many years, perhaps as a standalone subscription or as part of Microsoft 365. I feel that the updating regime is too chaotic and too frequent, a horrible combination. There were too many years with no vision at all, and there’s not enough information right now for anyone to judge the current strategy. At least there is one.

Regarding the most likely ways in which Microsoft might make Windows 11 “better,” which is perhaps subjective, nothing stands out. Apple is reportedly pulling a Snow Leopard with its next generation of OSes, meaning it will focus on cleaning things up, fixing bugs, and the like, and Windows 11 could use a year like that. Windows 11 has a (subjectively) beautiful UI in WinUI 3, but it’s also buggy and slow as death, and that needs to be fixed. (Less important and less likely, the UI inconsistencies endure in Windows 11, of course, and it would be nice to see that fixed too.) I would love to see Microsoft slow the new feature train down to a biannual or quarterly schedule, but we all know that isn’t happening and the deluge of AI features both useful and pointless will simply continue.

The only sliver of hope in all this is that Microsoft seems to have finally gotten the opt-in message on AI in Windows, and it does allow users to turn off basically anything AI related and most of the truly bad behaviors. The worry is that this stalemate is temporary, and perhaps tied to the speed at which AI moves, and so even Microsoft doesn’t know where it wants to be. And that sometime in the future, AI stuff, or, more importantly, the truly bad behaviors will no longer be opt-in and/or fixable.

For now at least, I’m uneasy but OK with where things are. We can fix the bad things. Windows is still better than macOS, Linux, or whatever else … for now. But this coming year will see Aluminum join the iPad as a device-based alternative to Windows and the PC on laptops. Linux is so much better than it’s ever been. And we’ll see. The truest thing I can say about my own experiences using Windows is that I would have left by now if there was a better alternative, and it’s only because there is not (at least for me) that I’m still here. That could change. And if Windows still sucks when/if it does, I am willing to walk away.

But again. I don’t think most people care. Windows and the PC are tools, they’re for work, and few people, relatively speaking, are enthusiastic about this stuff.

? Give a little bit, give a little bit of your life to me

Markld asks:

Loved your Little Tech Little AI post! So my ADHD, Lol, led me to this: It rings true that the most interesting breakthroughs now seem to be coming from smaller, more nimble players.

Yes. There have always been alternatives, but we live in a golden age of Little Tech alternatives. And that’s important because it opens up choice, and we don’t have to sacrifice quality, experience, cost, of anything else.

I’ve been thinking about who might make the biggest impact in 2026 if this “little tech / little AI” trend holds. I’d love to hear your thoughts on whether these kinds of companies can really shift the balance away, even slightly, from Big Tech, or if they eventually get absorbed by it.

I keep coming across additional companies and products that need to be added to the Little Tech “approved” list. Stardock, for example. Epic Games and GOG. It’s one of those things that grows and grows the more you think about it.

But here, I’ll take your list and comment on it company by company.

• Perplexity AI. Leaving aside the issue with IP/copyright theft, in part because it’s not unique in this space in that regard, many of the general purpose chatbots might be seen as Google Search alternatives, and it might be worth all of us looking more closely at those companies doing AI and search, like DuckDuckGo. But yes, Perplexity feels like a reasonable Google and OpenAI alternative, and its web browser is quite interesting. Trusting Perplexity is perhaps beyond me, but helping matters is that we can’t trust Google either.
• xAI (Grok). I will never use anything that’s tied to Elon Musk, sorry.
• Mistral AI. Corey Doctorow has begun talking about the “post-American Internet,” and European companies, especially those that are transparent and/or literally open source are key to that potential future. Europe is desperate to lessen its reliance on U.S. technology for understandable reasons, and I bet they’re not alone. So this kind of thing is super important, and as LLMs and SLMs get more capable, Mistral AI and AIs like that could emerge as interesting alternatives.
• Hugging Face. It’s possible I’m missing the point, but Hugging Face seems like a technical place to experiment with models today. It clearly wants to be more than that, and if you think about the BYOM (bring your own model) approach Brave is trying, or the various AIs that are anonymized front-ends for other (often bigger) AIs, then you see where Hugging Face could be going. It could become less of an aggregator and more of an orchestrator. That’s something an OS typically does, and it’s something we’ll see in web browsers, especially with the move to agentic capabilities. Is there room for a standalone version of that? Yes.
• Luma AI. I don’t have any experience with Luma AI yet, and hopefully no one is confusing it with Proton Lumo AI. But I had written something a while back called When Everything is AI, Nothing is AI (Premium) that I think explains Luma AI. Google, Microsoft, OpenAI, and others are pushing this “Big AI” thing right now, but the truth is, all these capabilities will be everywhere. And in the sense that many might use a free tool like Affinity or a reasonably inexpensive product like Adobe Photoshop/Premiere Elements instead of an expensive Adobe CC suite, many will turn to smaller and presumably less expensive AI solutions like this too. Again, golden age.
• Brave Software. Brave has the only web browser that anyone needs and a decent search engine, and for a lot of work, that’s enough. I would love to see them partner with other companies in this space or branch out into more areas. But that’s …
• Proton, GitLab, Supabase, Chainguard, DuckDuckGo, and Signal. I would add Notion to this list, with a bullet. And things like LibreOffice. And also Slack, sort of, though I hate it and I know it’s owned by Salesforce. The thing is, these companies, and many others, some noted above, can help form the basis of a DIY personal productivity tech stack, as I wrote this past April.

One of the tricks here is lessening our reliance on Big Tech cloud infrastructure specifically. I’ve done that with the two NAS devices, but I also continue paying for Microsoft 365 Family and Google Workspace/One, and use that storage too. But you have to take the first steps before you can run.

My mantra for 2026 is to move as much as possible to little tech – little AI.

This is my goal as well. Not to make a point, and not when doing so means I am losing out on something I need or want. But increasingly, I see myself using Little Tech over Big Tech. And we’ll see where it takes me. And us. I feel like I’m not alone on this “I’m mad as hell and I’m not going to take it anymore” thing.

? Site stuff

gg1 asks:

Any 2026 goals for the site you’d like to share?

I guess I’m sort of interchangeable with the site in some ways. And my overreaching goal is unchanged in that what I want to do is write, and have these conversations with you and others about the personal technology we all care about, and learn in the process. But this is a business, too, and I literally hate that side of it. The state of Pennsylvania requires me to submit an annual report, for example. There are taxes to deal with, and payment systems. Infrastructure to pay for that includes web hosting, CloudFlare, and whatever else. I wish I could ignore it. I try to, as much as possible.

Last year, we started the process of moving people over to a new membership system, I restarted the Thurrott Premium newsletter, and I started recording the Desk to Destination podcast with Stephen Rose. I was also delighted to see what Chris Hoffman came up with for his new newsletter, the Windows Readme. This year, I would like to expand on all of that.

The membership stuff is mostly behind the scenes, I guess, but the goal for 2026 is to move over fully. This is better for everyone, with self-service capabilities, reminders about upcoming payments so there are no surprises, and so on. But we’re also going to build out the points/rewards system with leaderboards and the like. So we’ll get going on that soon, by which I mean that Robert has been bugging me to get off my tush and get going on that, and I will.

I will likely have a new podcast soon and it’s one I think you’ll like. I don’t mean to be cryptic, but I haven’t finalized plans with my co-host and should do that soon.

I would also like to better budget my time this year regarding specific site-related activities. For example, while I’m not sure that I’ll have a focus each month, I will be focusing on security this month. I have various series in various points of progress, and I’d like to be more regimental about that, especially since I have dozens of partially completed articles I should finish up and get out in the world. And as noted above, I think a big overall focus this year will be on Little Tech in a pragmatic way.

I feel like the swag store needs more love, though that will never be a money maker.

There are books to write, including some I’ve never discussed.

I don’t know. There’s always so much to do and so little time. I feel like I never get caught up. But I’m not even sure what that would look like at this point. This is kind of the classic IT admin dilemma where you want to be proactive but life happens and you spend the year just reacting to things. (See the note above about monthly focuses.)

What would you like to see?

? Happy New Year!