• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

NVIDIA and Microsoft Reinvent Windows PCs (laptops) for the Age of Personal AI

I watched a video which said that the ARM CPU cores in the Spark are already 2 years old at this point. The rest of it is impressive but only for specific workloads. I think they're trying to collectively push AI further onto everyone when there is a general pushback and reluctance by the majority from more OS-level integration. Don't get me wrong, I wouldn't mind playing around with an LLM for development purposes, but I imagine I'm very much in the minority and I just don't think anything you do get will justify the cost. No doubt they'll eventually try and sell it to gamers with:

"Claw, beat this boss in this game for me"

....but that just defeats the point in playing the game if there's no challenge.
 
Last edited:
I dont see the point of local AI if you are not a coder? You dont need to memorise all the syntax and contexts for sql…

But for a normal windowser, any point in speaking to your pc, "windows, check neogaf, if any new thread on project helix, post a few concerned trolling replies in the span of 4 hours, tone is dry wit but with some technology terminologies "

We're at the innovators maybe borderline early adopters for this product

Product-Adoption-Curve-2-1-1536x759.png.webp


You don't see a use, the product's not for you then. You do not need 1 petaflop + 128 GB to ask it to open neogaf or to open steam 🤷‍♂️

The early adopters and development of the OS in time will make it that it'll become part of your daily life faster than you can think of.

Yes we can stopped that from happening! By working together to stamp out this AI bubble!

Cute

You ain't putting the genie back in its bottle. All the anti-AI effort will amount to nothing. Even if all customers in the world stopped buying it, it's not stopping what's going on behind the curtain. All governments and especially USA vs China will basically have a race to AI singularity to the likes of a project bigger than the Manhattan project + space race combined.

They will not stop. Giving the edge to another country means you'll be left in the dust technology wise.

Besides why would I need a computer to generate tokens so that it can then automate some stuff??? I just want to click and do it myself! Why would anyone want waste hardware on a middleman?

Why would I need internet when I can go and walk physically into the library and pick up a book?

Really fun to see some of those reactions. It's truely a perfect match to so many technology resistance case studies.
 
Last edited:
Yes we can stopped that from happening! By working together to stamp out this AI bubble!
AI is the engine of the American economy now. Otherwise it's healthcare admin spending and debt fueled financial services. If AI Blackhole erupts into an event horizon, you better pray to avoid the layoffs that will commence afterwards due to a 2008+++ crash.

AI is one of the greatest leaps in the history of mankind. Stop being a luddite.
 
Last edited:
No it's the same clang 15 compiler. Windows just sucks and nerfs GB6 scores for all ISAs by 5-7% compared with good operating systems like Linux, Android, IOS, MacOS, etc.

Oh I thought it was different but I guess not

Windows does carrie more general-purpose baggage by default so I am sure some of the difference comes from that as well on top of whatever the kernel does differently
 
We're at the innovators maybe borderline early adopters for this product

Product-Adoption-Curve-2-1-1536x759.png.webp


You don't see a use, the product's not for you then. You do not need 1 petaflop + 128 GB to ask it to open neogaf or to open steam 🤷‍♂️

The early adopters and development of the OS in time will make it that it'll become part of your daily life faster than you can think of.



Cute

You ain't putting the genie back in its bottle. All the anti-AI effort will amount to nothing. Even if all customers in the world stopped buying it, it's not stopping what's going on behind the curtain. All governments and especially USA vs China will basically have a race to AI singularity to the likes of a project bigger than the Manhattan project + space race combined.

They will not stop. Giving the edge to another country means you'll be left in the dust technology wise.



Why would I need internet when I can go and walk physically into the library and pick up a book?

Really fun to see some of those reactions. It's truely a perfect match to so many technology resistance case studies.
Oh my God Buggy Loop Buggy Loop that was perfectly written.

You know, most people don't understand this stuff, but I think you really have it in hand.

Early adopters in AI will have a huge leg up in the job market.

There is a reluctance by many, even here, to get into AI. They feel about AI how I felt about NFTs. Just another fly by night arbitrary tech.

People need to work a little more with it to understand how beneficial it can be. Just the simple act of snapping a photo and uploading that photo to chatgpt to ask a question about it. Once the average person is at that level of usage(which I have now gotten both of my 70+ year old parents) then adoption will take off.

People just don't think it has a use case(this is basically ignorance) combined with the fear they face from a job market or the pick pick pick they get from AI being mentioned 5000 times a day. I get it.

You are right of course, they can't actually slow this thing down no matter what they do(unlike most other things like nfts) and so they are actually doing measurable harm to themselves by not being willing to even discuss its usage.

Hate, distain, cynicism, and conspiracy theories are what AI haters have in abundance.

So for clawbot or a local AI, that is something that has been highly requested due to restrictions that have to be put on "available for everyone" AI. With local AI of course we will be able to be an actual agent. By having it on the PC it can do things mainstream AI can't, like watch youtube videos without the need for a transcript, or research things that don't allow bots on their site. The laptop contains the necessary human license being used by an AI product to actually make it agentic, so I can see myself being interested in this. AI are is getting MUCH better too with this last round of chatgpt updates.
czw2bcvx5jhe1.png
 
O god when does it stop



This is literally how a PC has been operating since its inception: you give a command and the PC does the thing.

Why is Jensen turning into an AI freak so badly? What he says there makes zero sense. It's so freaking stupid!!!
 
Zacfoldor Zacfoldor Good post

One case that comes to mind with my personal experience is at work. When someone a few months ago decided to use a transcript AI to make resume of the team meeting we had, like a dozen peoples, which was full of technical terms, an engineering project, I was fucking FLOORED how it resumed everything we said in concise points, the things that were agreed, the things that needed further development and pending questions.

Project managers used to take these notes manually, made mistakes, asked questions because they did not understand the technical words, painstaking process. Now AI takes the notes for him and he can actually focus more on what matters on his side.

Other use cases on the more technical side of the planning and processes were used but I'm not at the liberty to detail those but again, shocking results.

If I start a personal engineering project in the future to the likes of making something as optimized as possible to produce and performance (I did that in the past before AI), I would absolutely use AI over going through standard optimization routines like brute force, genetic algorithms, ant colony model, etc. Or at the very least, an hybrid system.

When I get home after working on a computer all day, I open the desktop only for entertainment. Youtube / neogaf / Steam pretty much. To peish peish 's point, me neither at home I don't need it, for what I'm doing, right now. Nor will I spend huge sums of money for these laptops or an apple M5 or a DGX Spark. If my field was in computer engineering it would likely be completely different. But I'm certain that it'll eventually just permeate everywhere and we'll be using it for daily tasks without realizing it.
 
Why would I need internet when I can go and walk physically into the library and pick up a book?

Really fun to see some of those reactions. It's truely a perfect match to so many technology resistance case studies.

walking to store or reading e-book is the same thing. One just more convenient you can download it anywhere.

I cant see where AI can come into this process that is helpful to me?

All that local agent AI will be just another process to compete for limited computing powers....50 chrome tabs is 50 chrome tabs taking up part of your memory and cpu. 🤷‍♀️

now you have another NPU layer taking up you memory ssd and battery life
 
No it's the same clang 15 compiler. Windows just sucks and nerfs GB6 scores for all ISAs by 5-7% compared with good operating systems like Linux, Android, IOS, MacOS, etc.

Wait, how? Could Windows even fix it without breaking backwards compatibility?
 
Last edited:
walking to store or reading e-book is the same thing. One just more convenient you can download it anywhere.

I cant see where AI can come into this process that is helpful to me?

All that local agent AI will be just another process to compete for limited computing powers....50 chrome tabs is 50 chrome tabs taking up part of your memory and cpu. 🤷‍♀️

now you have another NPU layer taking up you memory ssd and battery life

Read my post above where I detail using it for work

In the end, for my entertainment time at home when I open my gaming desktop, I also don't need AI. No I won't buy specifically an expensive high RAM high tops machine for AI work at home. I just know though eventually it'll probably become part of daily life and hardware will have advanced enough that you won't sell a kidney to have that, it's just gonna be there.

You don't need AI, that's fine, it's not gonna be a product that will sell any meaningful numbers, some would say DOA probably after pricing, but it's a start for the very enthusiastic local AI peoples. That's it.
 
Read my post above where I detail using it for work

In the end, for my entertainment time at home when I open my gaming desktop, I also don't need AI. No I won't buy specifically an expensive high RAM high tops machine for AI work at home. I just know though eventually it'll probably become part of daily life and hardware will have advanced enough that you won't sell a kidney to have that, it's just gonna be there.

You don't need AI, that's fine, it's not gonna be a product that will sell any meaningful numbers, some would say DOA probably after pricing, but it's a start for the very enthusiastic local AI peoples. That's it.

Exactly right. Not everyone is going to have interests in this stuff as it doesn't apply to them. For those of us whose jobs are directly impacted by all of this, it really isn't a choice It is either jump in or get left behind. That's how I look at it anyway so I've jumped onboard.

were the millers no ragrets GIF
 
Exactly right. Not everyone is going to have interests in this stuff as it doesn't apply to them. For those of us whose jobs are directly impacted by all of this, it really isn't a choice It is either jump in or get left behind. That's how I look at it anyway so I've jumped onboard.

were the millers no ragrets GIF

Option C: Get left behind and laugh at all the jerks who jumped on board. See: The Hindenburg.
 


If that's real, at that price for 128 GB it's really "cheap". I'm surprised. The M5 pro with 64GB is $3449. M5 max with 128 GB is $5549. Just the fucking adder for 128GB on Apple side is a grand.

You basically get a $4700 DGX Spark for $2899 ?

edit - oh. "needs" to be priced at, so that's just their calculated bare minimum? Yea I would expect higher prices when they unveil.
 
Last edited:
> RTX Spark

Basically the DGX: mobile and DGX: electric boogalo.
Brb, selling my kidney to run dense models like Qwen 3.6 27B / Gemma 4 31B at miserable speed with no tensor parallelism, quantized into 4-bit lobotomites.

inb4 GAF begins to defend it once again, like when someone replied in one of those threads - "you can't stack 3090s cause that's not how LLMs work" - man, that was just...
Is THIS the future of personal computing the people really want? I know some claim it's useful for "CUDA developers", but why the rest must suffer?
IF the cost wasn't so high, then yes, perhaps it wouldn't seem so crazy to buy such a device. Realistically, though, it's a step toward herding us into the cloud computing prison.


MBvQy6BgZcOihrrq.jpg

Nvfp4 isn't lobotomised, just optimised for Blackwell. I can run a dense model like qwen3.6-27b with a 256k context window at around 46tps 24x7.

Can lower the kv cache and allow more concurrency if I want

Source? I have a dgx spark and a 5090
 
Last edited:
Top Bottom