Ask Paul: June 26 ⭐

Happy Friday! We’re awash in a component crisis of confidence, given all the news this week. That makes sense, but there’s good with the bad, and some terrific reader questions to wade through.

📉 Margin call

jrzoomer asks:

Do you think Valve is going for Apple-like margins with the new Steam Machine? Gabe Newell with his new $500 million yacht and $70 million mansion seems to be gunning for that premium brand. And maybe this is a great setup for the upcoming XBox Helix?

I can’t imagine Valve would sell the Steam Machine under cost, since it’s not going to drive game sales or whatever; the audience for Steam Machine is literally dedicated Steam customers. Plus, Valve is in a Microsoft-like buying position when it comes to component pricing and availability. This thing is low-volume by design.

But if this week taught us anything, it’s that even the richest and most well-positioned hardware companies, like Apple, cannot escape the exploding costs of components during this crisis. Microsoft had already announced super-expensive new Surface laptops, and once Apple dropped its pricing bomb, XBOX quickly followed up with massive new price hikes of its own. The XBOX Helix will be even more expensive, I bet, unless some miracle happens.

(The Apple and XBOX price increases could help drive sales this summer. Apple didn’t include the iPhone in the price hikes, so that will come in September, meaning the current iPhones are as cheap as they’ll ever be. And the XBOX price hikes goes into effect in August.)

I mentioned this on Windows Weekly, but several friends reached out to me when Valve announced the Steam Machine pricing and availability, and each was shocked the price was so high. I can’t explain this, but when I saw the $1050 starting price, I was surprisingly OK with it (aside from the lack of a controller, which I found out about later). Given the component crisis, Valve’s lack of buying power, and that no one would have been surprised if Valve just canceled the thing and blamed the crisis, I’m surprised it happened when it did, let alone at all.

I do see the attraction here, but now that I’ve been gaming on laptops, I don’t personally want a game console that can only be used on a single screen in a single room in a single house. It just feels old-fashioned to me (like “a computer on every desk” went from visionary to quaint), and while I understand that big boxes will deliver the best visuals and performance overall, this is the right compromise for me. The audience for a dedicated console is small compared to other personal computing devices, and this is a niche within a niche.

Gabe must be doing something right. But from the outside, I don’t quite understand why Valve and Steam are so successful, and I don’t see them as a top-tier game developer. The work they’re doing on Proton and whatever else to make Linux play Windows games well is incredible, but it’s also emulation and I sometimes wonder if the better effort would be convincing game makers to port natively to Linux. Perhaps they did try that. But given the weirdness of Gabe and the way this company works, I suspect that Proton came out of some skunkworks thing an individual was working on and it just happened. I picture them all sleepwalking into their success.

“Siri AI is both impressive and disappointing”

In the good news department, that headline is just disappointing

💵 It pays to pay attention

wmurd118 asks:

I thought I was halfway up on tech, but I guess I’m pretty naïve. I’ve been reading and listening to all these stories about memory cost and I was freaked out about having to get a new PC. I’m retired and at my age, I only use my PC, an eight year-old Dell, about once or twice a month to keep it updated and to do a couple of other things. Most of my work I do on my iPad or my iPhone. I wanted to upgrade my Dell to 16 GB of RAM and I thought it was gonna cost about $500 to do that. Couldn’t see any reason to spend $500 when I could get a new PC or Apple for about $1000. So before I did anything, I launched Gemini and talked to it about what I wanted to do and it told me that I was thinking of new memory prices. It said the old memory, which is what I have in my old Dell PC, is significantly less because it’s old tech and companies have a lot in storage. So I went ahead and priced it out and for less than $90 I got 16 gig of RAM that I can plug-in and that will probably do me for a long time to come if not always. Could probably have gotten it for less than that, but I was satisfied with the Amazon price. Hopefully, most people out there, I wasn’t one of them, are aware that the shock stories about memory prices are for new memory rather than old memory. $90 versus $1000 is a good deal in my book.

This is a fascinating trend that’s tied to the discussions we’ve had about the silver linings of this crisis. Platform makers like Apple and Microsoft are digging into their OSes to make them more resource-friendly and efficient. Similar work is being done on apps. Users are holding onto older PCs and devices for longer than ever, and they will benefit from all the efficiency work. Etc.

On the hardware end, we’ve written about Intel and Qualcomm releasing lower-end new chips for inexpensive new PCs that can better compete with the MacBook Neo. But I don’t think we mentioned that AMD is keeping older CPUs in market too, and, in at least one case, bringing back older CPUs that are cheaper to produce and use lower-end RAM that inexpensive and readily available.

I love that you were able to upgrade your PC. But one of the downsides of all the advances we’ve seen to CPUs, chipsets, and PCs over the past decade or more is that we went away from user-replaceable RAM and other components in laptops in part to help make the devices thinner and lighter. The right to repair movement helped to reverse that to some degree, and most modern laptops are now user-serviceable and more easily repairable as a result. But the one thing still missing, in probably 95 percent of new (portable) PCs is user-serviceable RAM. It’s a huge problem.

Part of the issue is that some modern chipsets (Apple, Qualcomm) integrate RAM directly into the CPU on the SoC. Some (Qualcomm, AMD, Intel with Lunar Lake) integrate RAM into the SoC, and/or PC makers use soldered RAM because SO-DIMMs are too big and an added cost. But every once in a while, I will see a laptop with more modern CAMM (Compression Attached Memory Module) RAM, which can be user replaced or upgraded. I wish there was more of that: If you buy a MacBook Neo, an 8 GB Surface Pro/Laptop, and most other portable PCs, the RAM you get at purchase time is all you get, forever.

Anyway, this isn’t a “do more with less” situation if you have an existing PC and want to keep it going. It’s a “keep doing what you’re doing” situation, and one that can be improved if your PC is upgradeable and those upgrades are reasonable. This doesn’t excuse the stupidity of the AI-triggered RAM and component crisis, but it can soften the blow.

“CMF Phone 3 Pro officially cancelled for 2026”

The word officially is officially unnecessary in a headline

🔊 Not so smart speakers

wright_is asks:

You were talking about the new Google speakers and how they compare to the Apple and Amazon “equivalents”, although, as you say, these sit somewhere between the minis and full sized versions of the other brands.

I have a pair of HomePod minis that I use as bedside speakers for listening to audiobooks to go to sleep to. I was waiting for the new generation to come out to maybe extend them to using them with the Apple TV in the living room, but the delay in playback, as the audio is buffered to the speakers would drive my wife nuts and the things would be thrown out within a week!

Yep. There’s so much about the HomePod speakers I find infuriating, but the biggest for me is the paradox of an Apple product that does not just work out of the box, is next to impossible to update because there is literally no way to click a button to make it happen, and then just as impossible to troubleshoot when something goes wrong. I own four of these things, and while regret is a strong word, the state of this hardware is borderline shocking. Fans buy more Apple stuff because of the good experiences, and HomePods undermine that.

I was also looking at using something like that for my MacBooks (work and home), when I am in my home office, but I ended up getting some simple, and relatively cheap Edifier speakers with USB-C connection.

Edifier is an excellent choice. I still have a pair of Edifier powered bookshelf monitors and they’re almost as impressive as my Sonos Play 5s, sound-wise. But they don’t have any modern/smart home connectivity, so I’d have to add that with a dongle or adapter or whatever. Which I lurch into researching from time-to-time. This was driven in part by Sonos problems big and small, not just Sonosgate, but also its antitrust-related breakup with Google, which makes its products less useful for Android users, and other weird events like its breakup with IKEA, which was making super-inexpensive Sonos hardware.

About a year and a half ago, I wrote Looking Past Sonos (Premium), which was, perhaps, premature. Coincidental to the component crisis that keeps coming up now, I have a lot of Sonos equipment that just works, so it makes sense to just keep using it, given the expense of replacing it and the complexity of choosing something that meets all my needs. In my case, I’m an audio enthusiast, I guess, but not an audiophile, and not being rich, my choices are nicely limited. What I am looking for is something like Sonos that works with iPhone (AirPlay) and Android (Google Cast), offers solid sound quality, and can handle whole-house audio. In other words, what I want is what Sonos was before the Google breakup.

What I’ve reconciled for now is to keep using what I have while researching alternatives and paying attention to related announcements (like this one from Bose, though there are many others). And that brings me to an interesting coincidence…

Do you have any recommendations for smart speakers? Do they actually make sense? Or are proper speakers with streaming, or USB-C or TOSlink connected speakers a better bet? I sort of see networked speakers like a smartTV, you can use them as smart devices for a couple of years, but then the security updates stop coming and you have to remove them from the network, whereas the speaker side, generally, will last for 20 years or more …

This will vary by individual, their needs/wants, how one listens to music and other content, and so on. Smart speakers are convenient and usually inexpensive (at least compared to more traditional setups with an amp, passive speakers, and other components), so they match well with the friction-free worlds of Spotify, Apple TV, and whatever else. Sonos and its competitors offer an agnostic, higher-end smart speaker experience that combines some of the convenience of first-party smart speakers with some of the audio quality of enthusiast/audiophile setups. You could go nuts and spend a lot of money on components and speakers, which is not convenient and will require a lot of research. Perhaps most intriguingly, you can mix and match too.

And on that note, the coincidence is that I just spent a good chunk of time going down a rabbit hole of YouTube content related to WiiM audio streamers and smart speakers that took a turn into content about audio CDs in the modern era and whatever else. (For whatever it’s worth, I just found this guy on YouTube, watched a dozen or so of his videos, and find the content quite interesting.) I wasn’t looking for this stuff, though I am, in fact, blocking out some An Inconvenient Truth posts tied to me killing subscription services, and one of them will focus on music and handling the local audio/streaming service divide.

From a hardware perspective, there are a lot of cost and convenience issues to consider. You mention longevity, which is understandable and true, as passive speakers will outlive us all and let owners upgrade other components all around them. But there’s a lot of complexity there too. It all depends on where you land on that spectrum or matrix of cost, convenience, complexity, compatibility, typical usage, and whatever else.

I guess I’m in the middle somewhere. I love music. Sound quality matters, and the Sonos speakers I have are terrific and sound great, but they don’t have Dolby Atmos. (My HomePods do, but this is barely noticeable with a stereo pair, so whatever.) It just working really matters to me, as I gripe when things go wrong, but also because I have a wife and she’s not taking a class to figure out how the speakers work. And then I use whatever services and sources for music and whatever I use has to work with all that.

As noted, I’m going to stick with what I have and for all the right reasons. If I had to replace all my speakers for whatever made-up reason, I would need to resolve a few use cases:

• Home theater, which is currently a stereo pair (here and in Mexico) of speakers. But this could be a soundbar (part of a broader whole-house ecosystem or standalone) or a better Dolby Atmos multi-speaker setup with rear speakers and whatever else. Basically, this is about whether the home theater components (whatever combination of soundbar, speakers, and sub) are specific to the TV or part of a whole-house system like Sonos.
• Living room, which today is the TV setup plus other speakers in the room and nearby.
• Elsewhere in the house, like the bathroom, where I’ve alternated between a single Sonos or Bluetooth speaker so I can listen to audiobooks, podcasts, YouTube videos, music, or whatever else while shaving and showering. We also have speakers in the bedroom and kitchen, where it’s nice to have content going when doing the dishes or cleaning the house of whatever.
• On the go, which is usually going to center on apps on the phone and maybe tie into CarPlay/Android Auto.

It’s daunting. I like that the Sonos soundbar, sub, and speakers in the living room work with the TV (on an Apple TV) but can also be targeted individually if we’re just playing music. There’s a flexibility there that will be difficult (and/or expensive) to duplicate with another smart speaker setup or some component setup with passive speakers. I couldn’t affording wiping and replacing the whole setup at once, too, so I something modular or extensible is important. If this thing future-proof?

Here in Pennsylvania, my current mindset would be a solid component-based system with passive speakers and some kind of WiiM or similar streamer. In Mexico, which is tiny, I would probably just stick to two smart speakers on the Apple TV, probably Sonos Era 100 SLs (with no mics) now that you can use any speakers as the default on Apple TV. But I could also see using a WiiM with an amp and a pair of inexpensive passive speakers there, too, I guess.

This is not easy. Things change, we’re stuck in some cases with previous decisions, and our needs all vary. But I think we’re on the same page with a lot of this. Open systems are better in general, rather than getting stuck with whatever lock-in, but convenience matters too. It should just work and not require constant adjustments or fiddling. Sound quality matters. Having speakers where you want them matters, and having them all work together might matter.

The Google Home Speaker I just got is OK. It’s Google-specific, of course, and two of these with a Google TV Streamer (or the Walmart Onn equivalent) would work as well in our small Mexico City apartment as do the Apple TV and HomePods Minis there now, I bet. But any Sonos speaker would be better from a sound quality perspective. And while I understand anyone instantly crossing Sonos off the list, I’ve come around to the company improving and trying to be better. There are Sonos-like alternatives that offer cross-ecosystem support and a variety of price points. Bose, Denon, Wiim, and, at the very high-end, Bluesound. And others, I’m sure. But I feel like a Wiim streaming box of whatever kind with some speakers would get the job done too. And that’s sort of where I’m leaning. Or would be, if I was switching right now. A balance between a full-on smart speaker system and more traditional components.

Sorry if this is all over the map. I have been thinking about this topic a lot lately tied to that subscription service/An Inconvenient Truth stuff. And as noted, I just coincidentally ran into some helpful YouTube content that got my mind racing as well. I wish this was obvious.

“Apple iPhone 18 Pro reportedly set for September launch”

Unofficially

💻 My WinUI ennui

Todd Sizemore asks:

Have you been exploring any WinUI 3 apps? I found out about Dowstofon by random chance and thought you might have an interest in it.

I wasn’t aware of this app (a Mastodon client), thanks.

I wouldn’t say I specifically seek out WinUI apps. When I look for apps in the Microsoft Store or elsewhere, I guess I am looking for things that look modern in Windows 11, and those will usually be WinUI-based. But I would put up with an old-school UI if that app was exceptional, too. And I am curious to see what Microsoft comes up with in its native app push, of course.

Longer term do you think Microsoft’s newfound focus on WinUI will net any new native apps by the community? Something akin to how Apple has a cadre of loyal indie app developers.

I’m a bit mixed on this.

WinUI–by which I mean the Windows App SDK plus WinUI 3–is a nightmare, and it is lacking in all kinds of ways. If you’re coming from a previous Microsoft framework, most likely WPF, it will be familiar on the surface, with C#, XAML, and similar APIs. But it’s also much more complex and missing an incredible amount of obvious functionality, and bridging that gap is difficult. And compared to modern frameworks from Apple (SwiftUI) and Google (Jetpack Compose, Flutter), or even React/React Native, WinUI is obsolete and old-fashioned. If you want that kind of experience in Windows as a developer, you have to use a compatible cross-platform framework (like Flutter).

What’s interesting is that the Windows Developer Skills set of agents and skills that Microsoft announced at Build 2026 can help overcome those complexities and help developers or technical users create WinUI apps pretty quickly using whatever AI and natural language. You could argue that this puts native Windows development, or at least vibe-coding, on par with what’s available from Apple and Google for their frameworks, as they both have similar tools and capabilities.

And that’s true, and fair, but because that WinUI code is so terrible and, for many, unfamiliar, there is a code maintenance issue here that I don’t normally think or care much about for personal vibe-coded apps that is unique to WinUI. And that’s just that more people understand SwiftUI and Jetpack Compose, so the apps one creates that way will be more easily debugged and fixed if something goes wrong.

That may not matter in time. If Microsoft can convince developers to create WinUI apps–either from scratch or starting with whatever vibe-coding method–then they will gain expertise in the framework and its idiosyncrasies. And Microsoft has (once again) promised to “fix” WinUI, which is some combination of open sourcing the Windows App SDK, adding missing functionality, addressing bugs and performance issues, and, maybe, just maybe, allowing WinUI developers to create app UIs using declarative C# code (instead of or in addition to XAML) via Microsoft UI Reactor.

But I keep coming back to the same issue. Why would any developer do this?

Yes, Windows 11 has a billion users, but they’re not engaged with finding new apps, they’re using the same legacy desktop apps they’ve used for years. I suspect that most developers who adopt WinUI will do so to modernize existing apps. And that it still makes the most sense for new apps to be cross-platform because other platforms will see much more usage. We should just be happy to get those apps on Windows at all.

You can create modern-looking apps on Windows using Flutter, React Native (which does use WinUI for the user interface), Uno, and more, and these things aren’t locked just to Windows. That they are, or are not, WinUI literally isn’t as important to me that they look and work great. And they can.

Logically speaking, Microsoft should never make yet another app framework for Windows, and the company explicitly said it has no plans to do so at Build. So modernizing the Windows App SDK, the app framework partner of the WinUI user interface framework does make sense. As does Microsoft actually using these technologies to make their own apps, whether they’re in-box in Windows 11 or not, at least to some degree. Whether third-party developers (especially notable developers) ever adopt that is an open question. But generally speaking, I feel like that ship has sailed.

“Choosing the right AI model in Microsoft 365: flexibility, control, and confidence”

We give up. You figure it out

🤬 Comment spammers

Simon asks:

The annoying forum spammers are back, the comments seem very obvious. Has your provider for spam protection looked into why they are getting through? On a related note, if we see spam, is there a good way to mark it as spam or just in any way for your attention?

So, this was something that OpenWeb did very well. And now that we’re back on stock WordPress, I am reminded that this is something WordPress does not do well at all. Which is part of the reason we used OpenWeb.

What I’m doing now is what I’ve always done, which is reading the comments throughout each day, now via a WordPress interface. When I see spam comments, I delete them and remove the user accounts from the system. This has gone from a once-a-week thing with OpenWeb to something I have to do once or twice each day usually. Because, again, OpenWeb was wonderful for this. And WordPress is not.

What I have been thinking about is changing how commenting (and forum posting) works.

Today, when a new user posts a comment with a hyperlink, it’s flagged and I have to moderate it (approve, mark as spam, or delete). And after that, comments use some kind of algorithm based on the age of the account and number of comments; for most of you, those comments will go live but they’re highlighted for me in the comments interface too. And the first time a new user posts to the forums, it’s moderated and I have to approve it, after which time future posts will go live automatically.

What I will likely ask Robert to do is change this so that no comments from new accounts go live, ever. I will have to moderate them first, and it’s usually obvious through some combination of what they commented and the newness of the user account (I can see the number and know that higher numbers of whatever range are new) that it’s just spam. So I can just delete the account.

Ideally, I would have some one-button answer to “no, this is spam, so kill this comment and the account,” but that may or may not be doable, and I’d have to pay for it.

What you can do is minimal right now, as there’s no obvious “mark as spam” system. I will ask Robert about that as well. Something where whatever number of spam reports hid a comment would be nice.

“Microsoft is turning Windows 11 into an AI operating system”

Oh, thanks for the heads-up, we’d never heard that before

🙏 Om

train_wreck asks:

Did you where Om Malik passed away? Big loss in the tech journalism space, condolences to his family of course but also to the wider tech press scene.

Yes. That’s a shame, he was obese and living poorly, to say the least, but he had a heart attack scare years ago and turned things around nicely. But it seems his past caught up with him. I didn’t even recognize him when I saw a recent photo. For those unfamiliar, he founded Gigaom 20 years ago and achieved levels of success I can barely comprehend.

“Intel’s Chip Business Shows Signs of Life After Years of Struggle”

No it doesn’t

💽 Turning on a dime

lvthunder asks:

Do you think the PC and device makers will band together (or do it alone) and finance the startup of any new memory and storage plants? Is this something Intel can do?

I was trying to think of an easy way to explain this, but calling it a “chicken and the egg” problem doesn’t quite ring true to me. It’s more of a “capacity, time, and market dynamics” problem, I guess.

It’s complicated. And I can’t claim to be an expert in this area, so bear with me.

Excluding some of the earlier conversations–like using older technology that’s not impacted by the component crisis–there’s no way for any company, other than the companies already producing RAM, to make up for the current shortfall. And there is no good reason for those companies to change a thing: Their fabs are already running at capacity and they can command any price they want.

Creating new fabs, whether it’s existing players or new entrants, is a non-starter too. Each would cost tens of billions of dollars and it would take at least 3-5 years to build them and get them operating at sufficient yields. By which time the crisis will be over and that expensive fab you just built is designed to create a once inexpensive thing that is now a commodity there’s no real demand for anymore.

(This has happened before, repeatedly. I see this in reading recently about the past, for example, the reason the VIC-20 had 5 KB of RAM is that Commodore had a surplus of inexpensive SDRAM at a time when the DRAM they would have used was so expensive it would have made that computer prohibitively expensive to produce. That was in the very early 1980s, and when the C64 happened two years later, DRAM prices had come down. And the Commodore stopped making RAM.)

Looked at another way, let’s ignore the barrier to entry and imagine some company–like Intel–could spin up a RAM fab relatively quickly and get to market in time to help ease the component crisis. So it gets this going at a cost of many tens of billions, magically achieves the yield it needs, and then it’s time to start selling RAM. Why on earth would any company sell RAM cheaply when every other company is selling it at great expensive to AI companies? It’s a publicly held company and the best way to achieve shareholder value and dig out of its financial problems is to just reap the rewards. Intel’s CEO would be fired for not taking advantage of that. It’s not a charity.

These things feed on themselves. We don’t know how long the component crisis will last. We do know how long it would take, best case scenario or not, to spin up new capacity and how much that would cost. There’s no way to calculate the circumstances when or if that effort would ever pay off. But even if it magically did, then the companies would be stuck with expensive overcapacity, physical assets that are now losing value quicker than usual. So they move slowly or not at all. They’re meeting demand, but exceeding that would never pay off.

For whatever it’s worth, Intel used to make RAM. I think that was its first big product, before it had successful microprocessors, but whatever. It exited that market many years ago for good reasons, and there were and still are other companies that make RAM (and storage) that works better and is produced less expensively than what Intel could/can do. Intel can’t even bring new microprocessor fabs online quickly, especially not for modern process technologies they can never seem to perfect, and that’s what they do. RAM is a different product and a different market. They’re not vertically integrated that way by design, but even if Intel wanted to make high-quality RAM at scale, we’re looking at many years and too many billions of dollars. It’s impossible.

I keep thinking that the component crisis has to pass at some point. Richard has told me a few times that he sees a theoretical future of component glut and low prices, and that we’ll all be using PCs with 128 GB of RAM in a few short years. But I’m not so sure about that, despite history, and in part because AI skews everything so much. It’s possible that the component crisis simply eases over time, and that prices will come down only gradually and could plateau well above pre-crisis days. That this is in some ways a “new normal.” This is worrying, but it feels likely to me. We were all living on borrowed time and just didn’t know it.

One more thought, and it’s somewhat contrary to the above.

Given Apple’s history, its desire to aggressively drop partners and do things in-house, its success with Apple Silicon, and the fact that those chips are successful in part because of their integrated RAM, one could make the case that this company should make its own RAM. You know they looked at this before and during the crisis, and we can guess they’re not doing it because of the prohibitive costs, etc., the barrier to entry. But no one wants to be beholden to others, let alone the whims of others that have little to do with them. And this is an interesting case where I could see them doing down this path.

Doing so would not help the rest of the industry, and Apple is still beholden to the same cost and time issues as any other company. But I can see it. I wouldn’t be surprised to find out this effort is already underway. That this is a “never again” scenario for them, having just been forced to raise prices so dramatically. Apple is unique, so this might make sense for them.

Maybe.