311–320 of 554 posts
It is a weird product for the Framework brand. The pitch for the Framework laptop is that it is repairable/upgradable/modular. Something that is uncommon for laptops nowadays. This is the opposite. Desktops are modular by default, so much is that my computer is like the Ship of Theseus, I never changed it, but upgrade to upgrade, it is a completely different machine than it once was (it started off as a 486!). This o…
Earlier quoted context omitted.
I know very few people that do anything other than upgrade their GPU or SSD during the entire lifespan of their computer. Maybe when I was younger I'd upgrade the RAM after saving up, but am fortunate enough now to be able to buy what I want up front. This product is for me. A few years ago I tried to repurpose a desktop with a bad motherboard, but it was impossible to find a replacement for the 7-year-old CPU. eBay…
Something I've noticed over the years is that a lot of PC enthusiast discussion seems to be self-selecting for those most likely to chase the latest hardware, which affects how they think and talk about future proofing or upgrade ability. The challenge with x86 PC is that because the platform is so flexible it casts the widest net over huge amounts of use cases and circumstances. The example that comes to mind is cri…
That 64gb OCed 3200 (maybe 3500) mHz DDR3 RAM for example, kept up with DDR4 speeds for most of that decade when I benchmarked every few years. Intel extreme processor, workstation motherboard, and GTX Titans also kept chugging along. I’ve been looking at some of the higher end DDR5 coming down the pipeline and those are also likely to be on the higher end of performance for a decade while the consumer parts catch up. The only problem is that the workstation motherboard market seems to have disappeared.
(This was a work computer for my consulting)
Earlier quoted context omitted.
They went for local LLM route and for that high-bandwidth memory is a must. Consider it a low-price alternative to mac mini or nvidia’s box. This can also be chained though not as effectively as macs for example (those have thunderbolt for interconnect)
> low-price alternative to mac mini This starts at 1099$ per the article, so the mac mini is the low-price alternative here.
Earlier quoted context omitted.
> low-price alternative to mac mini This starts at 1099$ per the article, so the mac mini is the low-price alternative here.
If you can find me a Mac Mini with 32GB of RAM I'd love to see it.
Select 32
It's not exactly an out of the park win though as it's only $100-$200 cheaper and it trends poorly the farther from comparing base models you go (particularly after 64 GB, for which you need to leave the Mini family for the Studio). By the time you get to 128 GB, which you'd want for the 70B class AI models, you end up back at the original statement.
Earlier quoted context omitted.
They went for local LLM route and for that high-bandwidth memory is a must. Consider it a low-price alternative to mac mini or nvidia’s box. This can also be chained though not as effectively as macs for example (those have thunderbolt for interconnect)
> low-price alternative to mac mini This starts at 1099$ per the article, so the mac mini is the low-price alternative here.
Earlier quoted context omitted.
To the best of my knowledge the ARM ecosystem is an absolute pain to work in, you can get Phone/Tablet SoCs painfully encumbered with out of date drivers and binary blobs. Or you can get enormous server processors that will cost $1000+. There just isn't much that's suitable for making a desktop or laptop that would meet Frameworks markets expectations.
what about the Snapdragon? Microsoft is using in their ARM laptops
Earlier quoted context omitted.
It's likely chosen for cost. There's a couple of brand new fairly cheap laptops with exactly the same screens on paper and a few other similar sized laptops that are in the ballpark. Skimming Google, there are pretty much are no laptops 12" and higher resolution than 1080x1200 that's current nor made by Apple.
On sibling thread I already mention the Surface Pro (convertible) at 12'' and it's 2880 x 1920. The next 2025 convertible that I found, Latitude 7350, is also 2880 x 1920 (at 13'', though). In fact, most of the 12'' convertibles with 1080p are either sub$800 (which I doubt this thing is) or come from Lenovo (whom you really do NOT want to compare with regarding screen quality -- https://www.notebookcheck.net/Enough-w…
But yes, considering that Framework's Chromebook is/was expensive (over $1k) for the class doesn't give a lot of faith that the 12in model would be any different. Though that was equipped with a 2k screen and had a 12th gen i5, so I genuinely wonder what kind of meetings happened to ship out a premium Chromebook (a segment that does not exist)
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-...
I don’t quite get why framework focuses so much on Intel and AMD. ARM laptops are in the rise, and don’t need active cooling. It’s hard for me to think of upgrading to another laptop with fans when so many fanless (I.e.: silent) options are available.
Earlier quoted context omitted.
I don't know anything about RAMs or their bus size. Is this something that will be "fixed" in the future, idk, with DDR6? Meaning we can have replacable RAM with such bus.
There isn't any standard to put that kind of RAM on DIMM-like slots.
Earlier quoted context omitted.
The soldered RAM is surprising for Framework, and doubly surprising for being so in a form-factor that usually doesn't have soldered RAM. Similar to what other commenters have expressed, it just seems like they shouldn't have built this product if they couldn't figure out the soldered RAM bit.
Apparently this is due to signal integrity. AMD says swappable RAM is not possible. Source: LinusTechTips video.
If LPCAMM already can't keep up with requirements when it is barely even out, then my guess is it won't fare well going forward.
So 1 of 2 things is probably true:
AMD is not being completely truthful with their statements that LPCAMM wasn't able to work (maybe it was just more difficult/complicated than they were willing to do, but it could work or
latency/speed requirements have already outpaced what LPCAMM can provide and soldered ram is the future.
I really hope it's the former, but it wouldn't be the first time something like the second has occurred. Apparently cache also used to be a separate, swappable component before it became integrated into the die. RAM might end up going the same way.