Live data from Hacker News

Framework's first desktop is a strange–but unique–mini ITX gaming PC

arstechnica.com

131–140 of 554 posts

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#131

Earlier quoted context omitted.

Yet another 1080p garbage screen. Please! I had tablets with higher DPI 10 years ago!

1920x1200 / 16:10. It's perfect for a 12" IMO.

No, it is not. It would be perfect for a 6'' phone, maybe. The goal is to have at least double the pixel density, and my 2016 tablet can reach this ( ~2700x1800 at 12'' ).

Even Surfaces have been using 1440p at 12'' since 2016, and 2880 x 1920 since 2018! Why would Android & Apple tablets at much smaller screen size have higher DPIs, if 1080p was perfect? Do you expect to put Android tablets closer to your face than x86 tablets for some reason?

Sigh... since when has DPI started _decreasing_ again? I refuse to accept this trend, in the same way it was stupid back in the 2000s when LCDs became a thing.

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#133

Earlier quoted context omitted.

I should still be able to click around the pages to browse the products, even if full cart functionality isn't there. Whoever is in charge of that website gets an L. For that matter, it makes me reconsider my hosting of things with Cloudflare. I know nothing about framework's site's configuration, but I know I don't want my site to have a waiting line like that. edit: also, the timer went down and then went back up,…

until the user logs in, the cart functionality is implanted client side with cookies, and incurs no db hit. guess it's true what they say about hardware vs software. you gotta pick one to be good at, and the other is going to suffer for it, to varying degrees. (inb4 someone mentions Apple. Apple is a hardware company. Their software's alright but it's full of bugs and they're simply not as good as it as they are with…

How about Nvidia, they have both good hardware/software.

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#134
post #14

Framework is great because they took an entshittified category and made a good and repairable product in it, upgradeable in a way the other vendors refused to enable. That was the laptop. Now they made a less repairable desktop PC. This brings nothing to the market. The soldered ram is particularly unacceptable. I get and believe that they could not make it work otherwise, but then they should have stopped the produc…

This isn't competing with normal desktops.

Better to think of it as a competitor to Mac Studio & Nvidia Digits, which are much less repairable by comparison. The soldered memory is an unfortunate reality of these "unified" memory systems.

The only way to get a traditional desktop with 96GB of VRAM is to spend upwards of $10K loading it up with 2-4 GPUs.

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#135
post #22

Two other Framework announcements: New 12-inch laptop form factor with 360 degree hinge (ie "tablet mode") and a touchscreen. No price announced, but it is aimed at students: https://arstechnica.com/gadgets/2025/02/frameworks-laptop-12... New mainboard upgrade options for Framework 13 models: https://arstechnica.com/gadgets/2025/02/framework-gives-its-...

Having seen how students treat school-provided Chromebooks, those IO modules will get lost and damaged at light speed.

That's what we call a canny business model!

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#136

Earlier quoted context omitted.

Here's hoping that touchscreen becomes available as a component for the 13 as well.

The 12 inch screen is a different aspect ratio so it would be unlikely.

Apparently it comes with an optional stylus, so there might indeed be a touchscreen there.

Their website was hugged to death, so I can't confirm.

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#137

Up to 128 GB RAM available for running AI models. Theoretically awesome, but this might have some interesting market consequences for everyone else.

The bandwidth still doesn't quite compare to a GPU, and 128GB doesn't fit DeepSeek R1, however. If they bump it to 512GB for $5000 or so, that will disrupt the market.

256GB/s is on par with the M4 Pro at a much lower price.

You could run R1 671B using unsloth’s quantized version that fits in <80GB. Not sure why that would be a benchmark though, there’s nothing that can run the model at full precision right now except for (very slow) server hardware.

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#138

Earlier quoted context omitted.

You can get a M3/M4 Max with 128 GB of RAM as well. The Studio will give you > 128 GB. I have a max with 64GB RAM, which is good enough for 70b models with a 3 bit quant. Even if I had more RAM to run larger models, my GPU would be the bottleneck.

> You can get a M3/M4 Max with 128 GB of RAM as well. To get an M4 Max so you can have 128GB, you need a macbook pro. The cheapest macbook pro with 128GB is $4700. M4 Max does have the benefit of more memory controllers, so it has twice as much memory bandwidth as Ryzen AI Max. But that's a lot of money to pay for it.

To get an M4 Max so you can have 128GB, you need a macbook pro.

By the time the Framework ships the Mac Studio will have been updated to the M4 Max. Although 128GB will still probably be around $3k.

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#139

Wait, so I’d have to use ROCm for compute? That sucks. I’ve had better luck with Intel’s drivers for their first series of dGPU’s. If this works with tinygrad’s AMD driver, that would then interest me.

My hope is that the popularity of this hardware creates pressure to improve ROCm software support.

Me & my 7900 XTX will be quite grateful if it does.

Re: Framework's first desktop is a strange–but unique–mini ITX gaming PC

#140
post #132

Is integrated/non-upgradeable RAM the unavoidable future for all consumer-grade devices? Is there any sort of standard on the horizon that would enable the massive memory bandwidth to compete against the M1-style approach?

LPCAMM2 memory is the answer, isn't it?

It's just not catching on, because of course it isn't. Manufacturers have no incentive to let you upgrade parts on your own.

Post reply on HN