Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Windows Copilot Runtime (windows.com)
82 points by plurby on May 21, 2024 | hide | past | favorite | 57 comments


12 cores, possibly all of them P-cores (they've said that SXE comes in bins up to 12 p-cores), and some early benchmarks are putting SXE cores as ~15% slower than M3? That indicates this is a $900 machine comparable to the entry-tier $2000 Mac Studio (12c/32gb), which is still on M2 and only has 8 p-cores. Its TDP is also pretty comparable to the M2 Max, at 80 watts.

This is quite an incredible achievement from Qualcomm. I'm very excited to get one of these and install Linux on it. And, stoked to see if future iterations of the Steam Deck eventually make the jump to ARM.


The new Qualcomm chips are the only thing I find exciting about all these recent Windows Copilot announcements. I'd love to get one of these new chips in a mini PC like the ones from GMKtek, e.g. https://www.gmktec.com/products/nucbox-g3-most-cost-effectiv...


I'm not sure if you've checked the link out, but that's what this article is announcing; a $900 Mac Mini-style devkit you can pre-order today.


Yeah, I saw that. The GMKtek boxes I linked to can fit in the palm of your hand, have great power efficiencies, and cost less than $150. I have a couple of the Intel N100 models but would love to get one with an ARM processor instead.


GMKtek? These boxes are inexpensive, but also are known sources of malware from China, no?


Ah - interesting, I hadn't seen that news. Appears that the Windows version they bundle with some PCs can be dodgy. I've never used Windows on these - I put in a new SSD and install Linux.


Firmware is also a vector. Particularly with Intel’s CPU remote management layers.

Maybe this wouldn’t affect ARM that much, that CPU is more obscure and with low volumes is a less of a target of interest (in the context of PC systems). But with more obscurity you wouldn’t even know where basic and most likely exploits can be based.


> I'm very excited to get one of these and install Linux on it.

If there ever was a time for Linux to really compete with Windows, that time would be now when Windows will break software compatibility to a large degree.

That really was Windows' big killer feature.


What compatibility is going to be broken?


I'm aware that there is a compatibility layer in place that will runs most apps, and that some of the most common production apps will have native builds.

But from what I can read, the compatibility layer is not perfect, and for instance games cannot be expected to just run or run well enough to be playable.

If you are at the point where Windows apps cannot be expected to just work, then why not use wine (and proton for games) instead, on an ad-free system?


> If you are at the point where Windows apps cannot be expected to just work, then why not use wine (and proton for games) instead, on an ad-free system?

Because they don't have to learn a new OS. I love Linux but inertia is a powerful force. I don't think swapping one compatibility layer for another is enough to entice the average person.

The ads, maybe, but I don't think most people care that much. Linux is a hard sell to people that don't want to care about compatibility layers, don't mind advertising that much, and grew up on Windows.


> I'm very excited to get one of these and install Linux on it.

Exactly everyone's immediate thought on this, lol. That the thing runs Windows is mostly a distraction.


Not my thought.

In general, Linux runs worse on laptops than Windows does. Many laptops have HiDPI displays, which Linux's graphics stack simply can't deal with properly. Track pads, fingerprint sensors, hybrid GPU set-ups, even external monitor arrangements are amongst the many issues Linux struggles with to this day. Linux has issues with pinning CPUs to a higher frequency than idle, and hence exhausting laptop batteries real quick.

I am loathe to see what having an entirely new hardware stack vis à vis Qualcomm will do to Linux installs on these.


I read a few articles about these, but I didn't see one that talked about the bootloader situation


It's UEFI and Debian works.


yeah i finally saw that and was pretty excited. We'll just have to see how they get locked down by the other OEMs


They use UEFI (like previous Snapdragons did) and EL2 is apparently unlocked on the SXE series (unlike previous Snapdragons.) It's mostly a matter of getting the device trees in place, I guess (I don't think these use ACPI natively, but might be wrong.)


"uses 20% ocean-bound-plastic"

Yep someone is literally standing next to a river with a net catching chip bag wrappers and drink bottles and sending them back to the computer factory.


What does that even mean though? Does it mean:

- 20% of the plastic in the entire device was recycled?

- 20% of the plastic in the box (as the context implies) was recycled?

- 20% of the recycled plastic was ocean-bound and the device is all or mostly recycled plastic?

- Is it some sort of 20% plastic-grade and the plastic was all ocean-bound?

- Did they capture ocean-bound plastic somewhere, used 20% for these boxes, and released the other 80% back into the ocean?

- Did they mean specifically "Ocean Bound Plastic" as a specific term (OBP) and not "ocean-bound-plastic" as a generic term? In that case, this website explains [0] that it would be 80% of all plastic that would end up in the ocean and that certifiable OBP is not commercially recyclable [1]. So... it's not OBP by definition then as OBP excludes all plastic that can be recycled commercially and this was?

- ???

When I clicked the 'Learn more' link next to it, the figure wasn't in the source page. There doesn't seem to be any source page for this, although a lot of news outlets are quoting it (in different interpretations). It seems like they were trying to greenwash something, though that article is wishy washy anyways.

[0] https://www.obpcert.org/what-is-ocean-bound-plastic-obp/ [1] https://www.obpcert.org/what-are-obp-credits/


It means they recycled the sprues from injection molding instead of throwing it in the river. About as truthful as carbon credits really.


Every year the new models will one-up the previous generation, using more and more ocean-bound plastic until someone gets paid to stand just upstream from the first guy throwing bottles and wrappers into the river to satisfy demand.


That… that would be pretty cool.


Remind me a story when French government in Vietnam started to pay bounty for rats tails


This is basically what the ocean cleanup project pivoted to.


20% ocean-bounded-plastic with 100% mussel smell. Double win.


ig it is better branding than "20% recycled plastic"


I'll wait until I see people test this to make any judgements, good or bad. I'm certainly interested to see what it looks like.


> New Source code integration in File Explorer allows tracking commit messages and file status directly in File Explorer.

And yet they make Home instead of My PC the default when opening Explorer, which then causes reads over the network if you happen to have network shares mapped and the whole thing to take several seconds to load. Microsoft has rather schizophrenic decision-making and zero will to fix things, they just keep building on top of the existing pile of garbage.


Weren't you always able to change the default Explorer folder? I just checked it on Windows 11, and here it's in Options (in the top bar, where the "New" button and file manipulation stuff is) -> Open File Explorer to -> This PC. But I'm almost certain this has been a thing since like, Windows 7?


Yes, you can change it, but the default is stupid.

> But I'm almost certain this has been a thing since like, Windows 7

Well, that only underscores my point on the zero will to fix things.


I vaguely remember fixing this in gpedit.msc but apparently it's even easier to fix now:

If you’d rather have File Explorer open to This PC, on the ribbon, select See more > Options > Open File Explorer to:. Then select This PC > Apply.


51 mentions of “AI” in less than 5000 words.

If anything, it’s infused in every layer of this press release.


bullshit.js really lights up the page.

https://mourner.github.io/bullshit.js/


This comment reminds me of the cloud to butt plugin that came out when everyone was making fun of "the cloud." Now you can't get a job in it without gcp, aws, or azure experience.


I still run cloud to butt, because it's still a stupid buzzword and the extension makes things funnier. Case in point: your comment reads to me as "my butt to butt plugin", which is hilarious.


> This is a bullshit leap in performance, made possible by a bullshit leap in bullshit.

> We look forward to continuing to bullshit with you, our developer and Bullshit community, to bring bullshit to our bullshit and tools, and bullshitting each of you to create future Bullshit bullshits that will bullshit every person on the planet

Is this a South Park episode script?


No, this is reality sadly.


“AI infused at every layer“

Okay… What does that even mean? And why does the company think we want that?

I love me some AI, but right now it’s getting put on everything like Nutella at a hipster eatery.


In practice it just means that the processor includes a matrix multiply unit somewhere.

For example, some of the latest Intel Xeon CPUs have 8x 1KB registers that can perform only one operation: matrix multiply (and accumulate results). See: https://en.wikipedia.org/wiki/Advanced_Matrix_Extensions

Generally this is very similar to the Vista-era Microsoft requirements for PC manufacturers, where it was a logo requirement to include a GPU that could be used to accelerate composition in the desktop window manager, such as transparency effects and blur. (Prior to that, low-end PCs had "graphics cards" that were fixed pipeline and not programmable /general purpose.)

Now Microsoft is forcing PC manufacturers to do the same kind of thing, but instead of a GPU it's now an "NPU" that they have to include. This can be a CPU instruction set, a co-processor, or a GPU baseline capability. The requirement is 40 teraoperations per second.

IMHO, 40 TOPS is way too low, and doesn't focus enough on memory bandwidth. Also, that 40 is total across CPU+GPU+NPU, which means in practice it'll require fiddly optimisation to get anywhere near that level of performance.

Windows Vista had the same issue, where many low-end laptops especially would struggle and spin up their GPU fans just from desktop workloads, let alone gaming...


"40 TOPS is way too low", I'm curious about this, from what I've read, Apple's M3 maxes out at 18 TOPS, and the newer M4 at 38 TOPS, so to me it sounds like even the entry-level Copilot+ PC is going to beat Apple's M3/M4 family. Am I misunderstanding.


It's possible that Microsoft's target is too low and Apple's performance is also too low. However when you find yourself saying that the entire industry is wrong you might want to stop and think.


An NVIDIA 4070 puts out 836 "AI TOPS" using 200 watts of power. So the 40 TOPS target is about 10 watts of power draw equivalent, assuming it uses a similar silicon logic tech to what NVIDIA used. With a more modern process, this is about 5 watts.

For an ultra-mobile tablet or laptop, this is... reasonable, I suppose.

For a desktop, it's quite a bit behind the current-gen tech, let alone "the future".

Like I was saying, this is aiming for where the puck was, not where the puck will be.

It's the same thing as Vista GPU requirements. Vendors will do the bare minimum that ticks the checkbox, but in practice it'll be useless garbage.

Modern AIs require > 1 TB/s memory bandwidth, > 128 GB memory capacity, and > 1,000 TOPS of compute to be really usable locally, not just technically capable of "running".


The problem is that "AI" isn't a feature. It's not a product. It's a pretty broad category of technologies that can be used to build features and products.

"AI infused at every layer" is like saying "we build our software with the best Agile(tm) practices". The customer doesn't care how you built the software, the customer cares what your software will do for them, and all "AI" means is "we probably have a deep learning inference engine somewhere under the hood, but no promises".


That's like saying audio playback isn't a feature or a product.


No, it's more like saying "media" isn't a feature or a product. It's not. It's a very vague word that could mean any number of loosely related things that do form part of many different features and products.


And that was totally a thing in the early 2000's

Companies were selling "Media PC's", and Windows ME stood for Millennium Edition or Media Edition depending on who you asked.

And people think the word "content" is overused now...


or, like saying "media ready"


I think it means many of us will not use Windows ever again.


I was already there when Windows 11 came along and blocked my perfectly functional 4 or 5 yr old HW from running it (even though the pre-releases ran just fine).

This "AI" bullshit is just another nail in the already nailed shut coffin.


Realistically, this is an announcement that Windows 12 will have a double click gguf runner and requisite backend as a feature.


Onnx, which seems to be awq


Most likely more effective data harvesting. Individual user level AI isn't the end goal I don't think. It's just a nice byproduct. This just means they can embed more data collection, and use your device to crunch the numbers for them.


This is the big thing I’m worried about with the AI revolution. If AI is being baked into your OS, where does usable training data end and your sensitive data start? Recall is pinkie promising that it won’t record sensitive data but at the same time it’s also saying “if your application doesn’t protect it well we mayyy see it and record it, whoops”. I can bet you 120% Recall is just a way for Microsoft to collect training data on millions of users every day. It is a privacy nightmare. But I don’t think the average consumer will care. Privacy died a long time ago.


We don’t want that. Investors want that.


it's a "fundamental transformation of the OS itself", you see


Probably true. The new win64 API will have just one entry point:

  LPSTR do_thing(LPSTR prompt);
E.g.:

  if (do_thing("get a mutex or whatever it's called") == do_thing("What's the happy return code for getting a mutex?") { ... }


Someone someday may actually ship code like this into production. Horrifying to think about. For some reason this reminds me of trying to grow plants with Brawndo https://www.youtube.com/watch?v=kAqIJZeeXEc




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: