It’s for AI acceleration. In AI training (and with LLMs, inference), the VRAM is basically hard limit of the complexity of the AI model you can run on the GPU. 20GB is enough to train some small LLMs, 10GB is not.

“Secure Encrypted Virtualization”

“Not for protection”

What the fuck is it ? Marketing???

I know about the displays. I’m talking about the difference on the GPU rendering literally anything 3D, which will be pretty massive.

Seriously considering this as my next primary work computer with me being a mixed AI/general software developer girl for everything. I’m in this weird place where day to do day 90% of the time I might just as well make do with a netbook because almost all my actual work lives remotely on workstations/servers, but once upon a time I can REALLY use scrunching a cool few tens of gigabytes of raw data locally without the latency of back and forth, especially to run some statistics on a dataset and to visualize it.

I currently have a Ryzen 5700X Lenovo Flex, which is effectively a Ryzen 2700X derivative, and while the CPU carries its weight pretty great, it was never offered with over 16GB of RAM and some data science garbage will just straight up run out of memory and wave goodbye and you can’t do anything about that, so the 32GB RAM alone would be nice, but I wouldn’t say no to 64GB either.

I’m also a sucker for OLED displays although 2944x1840 feels like a hella overkill, I wonder how well it would support halving the resolution; 1472x920 sounds like a really weird number but would probably go a long way towards making the battery life better and allowing you do some light 3D gaming.

The thing is this feels like the very limit of as good as laptops get these days without having a dGPU which I have zero use for; I do have desktops for that.

Can anyone else point out anything with comparable CPU and display but say, 64GB of RAM, but still without a dGPU?