Samsung Korea warns many apps won't run on its Qualcomm-powered Copilot+ PCs Samsung has debuted Copilot+ PCs running on Qualcomm processors, but warned buyers in South Korea – and seemingly only that nation – that they won't run many common applications. The term "Copilot+ PC" was dreamed up by Microsoft to denote a machine boasting a neural processing unit (NPU) capable of operating at 40 trillion …
Wrote a long article talking about all the problems based on what insiders at the OEMs were telling him. People seem to be skeptical since most of his articles are negative in tone and they wanted to believe Qualcomm and Microsoft's promises. But now that independent reviews are out, it seems like just about everything he was saying is being confirmed.
It's curious how people are automatically sceptical about anyone who is naysaying the current media push - as if they knew better and were more competent in deciding what was reliable and what is media bullshit.
That's why blockchain, virtual coin scams and AI are so successful. There are competent people raising problems, but nobody is interested. Especially not if "free money" is in the mix.
Until the pyramid crashes, that is, then everybody goes "well yeah, what did you expect ?".
Making one's own opinion is difficult, for sure, but I've always felt a healthy dose of scepticism is a good thing. Even more so when something is being touted as "the next best thing".
If it's true, I have time to see it become true.
"It's curious how people are automatically sceptical about anyone who is naysaying the current media push"
Not really. The human is a herd animal, so he or she who shouts loudest and longest gets believed and minority views cause discomfort, hence the drive for 'consensus' sought by committees.
It's worth considering that we used to burn folks alive for disagreeing about God, so skepticism about some current media push is a comparatively mild response.
Sometimes scathing is the *only* tone to take, when you are a lone voice against an orchestrated choir of praise !!!
This simply highlights that getting something 'out of the door' was more important than it *actually* working as MS & others pitch 'Eureka' !!!
This is a unholy mix of new architecture *with* new AI !!! .... [Where is the Holy Water and Rosary Beads when you need them !!!]
This AI, like all the others, is a con, based on *available* technology ....
Everytime, the *next* AI attempt is made to proclaimations of 'We have cracked it now !!!'
Everytime, the reality is nowhere near to the 'fever pitch' exclamations of how good it is !!!
For Gods sake, please don't get conned yet again ... keep your money in your pocket !!!
:)
It doesn't mean he's wrong, and because I haven't heard him speak, I can't say what I think about him specifically. However, the tone and the universality described can be a pretty big reason that someone might not be trusted. I'm thinking here of a reviewer of some niche devices I am interested in who can be counted on, whenever a new one is released, to tell you that it's crap and that nothing released after his favorite one from about 1997 was good. As it happens, I have used that one from 1997, and I wasn't that impressed. Several of the devices he reviews are actually crap, but because he approaches every review that way, the reader can't know whether this one is or not. If the reviewer doesn't try to be objective, including showing what things the subject does do well, and always treats the subject with negativity, then it can dissuade viewers from believing him or even reading the reviews.
I wouldn’t read it if he were always wrong. But it just takes a bit more effort than it needs to, sorting out the shouty bit from the valid insider info and analysis.
But it’s certainly true that he’s got a very high hit-rate about what’s going on inside Qualcomm; less reliable about ARM insider info. I have no clue (or interest TBH) whether he’s reliable on the Intel or AMD side. So yes, I’d have a high confidence that he’s called this story correctly.
For those who don’t know, here’s a very typical SemiAccurate story from the top
“Microsoft Throws Qualcomm Under The Bus Too
Exclusive: No one seems safe for long in the Windows space…. Microsoft seems to be on a roll of late stabbing all of their ‘partners’ in the back. SemiAccurate is actually shocked at the number of people who are actually shocked by their traditional corporate behavior….. So what are we talking about? No sooner did Microsoft throw the x86 players under the bus in favor of Qualcomm, they slammed said bus into reverse and ran over their new ‘valued partner”
I mean….is he right? I’m sure there is a relevant story. But I still know nothing at all about *what’s actually happened*, by the end of paragraph 4, other than Charlie is very derisive at some Qualcomm behaviour. Can I be bothered to read through that tone? Not always.
Well first of all you have to realize his site is subscription based, so his most recent lengthy article that wasn't behind a paywall was not the norm. He positions himself as the kind of subscription your work will pay for, it isn't priced like the NYT, more like Gartner. For some in the industry, I imagine it is worth it (after all he's been doing this for like a decade now so he's obviously got enough subscribers to make a living)
Second, if you look at what you can read in that post, and combine it with what we know about the failed Qualcomm X Elite launch, it would be quite understandable that Microsoft is pissed at Qualcomm for blowing their big "AI PC" on ARM party. Maybe there's a price to be paid for that, and maybe that price could be equated with Qualcomm being "thrown under the bus".
You might not like his tone, but you can't fault his ability to come up with scoops that no one else seemed to at the time like all the problems with these X Elite PCs, Nvidia's bumpgate issues with Apple, Intel's 10nm problems they denied over and over again, and so forth.
Well just guessing from having read his articles for a long time (dating back to The Inquirer - the UK tech website, not the US tabloid) I think he may figure there's enough fawning by the tech press so if he only has good things to say then it isn't writing because he'll just be echoing what others say.
He's built up industry contacts over the decades that allow him to get the inside scoop from people who trust him not to rat them out and jeopardize their job, or get off the record comments he can't publish but can use for further digging. You know, that "journalism" thing that's mostly extinct these days.
The abusive/scathing tone beyond the simple negativity may come from knowing that the people/companies involved could have done the right thing but chose lying/obfuscation instead. This is just another case of that pattern.
Chuck is pretty much a hero of tech journalism in my book, particularly for "Nvidia Bumpgate", going all the way to Electron Microscopy to confirm the correctness of his reporting, while under humongous pressure from business interests to shut him up. He sure rubbed some folks the wrong way in this (eg. biting the hand that feeds IT, at TheInq rather than ElReg) but definitely did the right thing from the standpoint of truth-in-journalism. Great to see him continuing that tradition at SemiAccurate.
The hope seems to be that expectations have changed. That with more/most people habitually using mobile devices they won't care about legacy device drivers and their old desktop software to the same extent. Whereas, as far as I can see, those are increasingly the only reasons for buying something with a "Windows" label.
Kind of. Windows RT prevented running apps not from the Microsoft Store. Windows 11 on Arm has x86 emulation and allows running anything. That said, the emulation might be too slow for games, or it might break cheat detection and DRM, or it might not be applicable to device drivers in kernel mode.
"Windows 11 on Arm has x86 emulation and allows running anything. That said, the emulation might be too slow for games, or it might break cheat detection and DRM, or it might not be applicable to device drivers in kernel mode."
You do recognize that your second sentence directly contradicts your first, don't you? Running "anything" means everything runs "as is". If that means you have to "emulate" by running a complete VM behind the scenes, then that's what you do. But if your emulator can't run everything "as is", it's called "incomplete" at best, although "broken" is more honest.
You know what was meant, though. It allows running a user-mode application. It doesn't allow device drivers. If the application has specifically written code (DRM) that doesn't work with it, that's not the fault of the emulation. Such things break in VMs all the time, or sometimes just because it was feeling like it one day. The main possible problem with this emulation is its speed.
I don't have firsthand experience with how good it is because I have not been convinced to want an ARM laptop. I have continued to assume that the emulation wouldn't be fast enough, even despite Apple's success with their similar emulation. I do know one person who has experience to the contrary, but they are running software that controls scientific equipment, which is probably not the most CPU-intensive program out there. Those who play a lot of games may have a different opinion.
But Apple have done it before, from Motorola to Power PC to Intel to AMD64.
Windows has in the past been available on DEC Alpha, MIPS, Power PC, and Itanic, but was never really usable on any of those platforms due to lack of 3rd party support.
The transition to AMD64 worked fine, because AMD64 supports Intel natively.
> But Apple have done it before, from Motorola to Power PC to Intel to AMD64.
For sure. All I meant is that (in Rosetta 2) we have an example of AMD64 architecture applications running very smoothly on ARM - so we know it can be done in principle, even if we have doubts about Microsoft doing it in practice.
Even Autodesk said that AMD64 AutoCAD Fusion ran faster on Apple M1 using Rosetta 2 than it did on an Intel Mac.
(Moot to my point is that Apple worked in advance to reduce the number of applications that required Rosetta 2 by steering developers to toolchains that could output ARM-native binaries)
(PS, is it possible you meant ARM when you wrote AMD64? It's ambiguous)
Userspace emulation is relatively simple and could be improved over time.
The major drawback is incompatibility with any x86 drivers, such as printer drivers or custom device drivers. These would not work, even in the future, unless the manufacturer make ARM drivers (which they probably won't — Brother has no plans for it for their current generation of printers for example).
Why licence anything when you can just use CUPS?
While I'd agree that CUPS has come a long way in the last quarter century, when last I looked, the CUPS ppds specified what print driver to use to process print requests for each printer it knows about. I doubt that just invoking CUPS will bypass the need for print drivers that run properly on ARM. I'd like to find I'm wrong about that, but my guess is that I'm not.
It's not just games and printers though, according to the Samsung "compatibility notice" linked by Simon (in Korean), and the article itself, the incompatibilities include such things as BNK Kyongnam Bank Personal Banking, Hyundai Insurance, FortiClient VPN, Adobe Illustrator, Google Drive, and more ...
It would benefit a lot of software companies if Microsoft could convince customers to break the legacy compatibility link with new Windows versions. Everyone benefits, software companies sell new versions of their products, the hardware manufacturers can stop updating drivers for products, and sell you a new thing. Extra double-plusgood is the opportunity to lock down new laptops so they only run signed MS OSes.
You users? Buy new stuff, or no computer for you!
I see it a bit different.
I don't mind buying new hardware for a genuinely useful new feature or speed increase. So for that its not a dig at me, and something Microsoft have tried a few times.
Often in these few times they had some.sort of an OS to go with the new hardware. Which again is fine in principal, happy to go along with it, after all I simultaneously lived with a version of windows, desktop Linux, mobile Linux, ps3 OS and previously Nokia. I am fine moving one to the other if there is a good reason to.
Where it ALWAYS falls down is the applications.
Which you will not are not something Microsoft have direct control over very often. They are often solely the care of various third parties. Those third parties are in business to make money off their products, which in some cases have and decades of updates to fix, tune and tweak every aspect of their programs.
Suddenly Microsoft launches new unseen hardware and is and promises the support will follow as they look across the stage at their software third parties and they are all just staring at their shoes pretending the building architecture is nice.
I'm not for one moment giving MS a pass on this, it is largely still a problem of their making by not working on this behind doors for as long as it takes to reach feature parity with x86. They develop announce and killed hardware and OSs faster than some application could be written for them.
What needs to happen is for MS to just take the 10 years or so and not talk it up, make a genuine new OS, ship it out to the hardware and software makers and let them work with it for years and get it working good. Then release in one big burst and smooth enough no one notices the change. Sort of like the entire apple approach, not that I like them or their stuff either, I just recognise it works well, when it works.
In the late 1990s South Korea mandated the use of Microsoft’s ActiveX for online services. Which was a very bad idea that came back to bite them. It could be a modern version of some such requirement, as governments never learn from their mistakes. SK has always been deeply protectionist in tech, hampering global services to promote locals ones. There may even be some state spookware involved. They are a bit paranoid about tech (esp. mapping apps) due to NK.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
