Hacker News new | ask | show | jobs
by npx 2437 days ago
I love this concept and I love that it is proliferating. Purism offers what I'd consider to be more attractive hardware, but that's very subjective. As of Kubuntu 19.10, I actually consider Linux to be the best operating system for a laptop.I look forward for 20.04 and a few years of near total disregard for updates.

I want to be an AMD fanboy! I want (at least) 16 cores, 32 threads, a fully open source graphics stack, Wayland, flicker free boot, and open source firmware as described here. As it is, I just had to go with Intel/NVidia because it's more seamless. Even though I'm not really using the NVidia GPU, I do have it available if I want to work with Tensorflow etc. Ultimately, for me it is a question of stability but I hope that these systems can really close the gap.

I want something like this System76 machine with very good support for encrypted ZFS right out of the box. I don't know if that would entail LUKS or ZFS encryption, but I want it to work. I want a USB key that actually serves as a key and allows me to boot or otherwise unlock the system. Again, I'd prefer this to be a fully open source AMD/ATI system based on Kubuntu. With ZFS, bpftrace, and Docker... this is what Solaris wanted to be when it grew up.

I'm not sure how big the market for this would be, but I'd pay good American money if anyone catered to it. Right now I'm using a Dell G3 Intel/Nvidia laptop which, in fairness, is obscenely fast.
1 comments
lonelappde 2437 days ago
If you aren't using a GPU regularly, Why not use a cloud gpu? Cloud CPU is cheap as free if it's just occasional playing around, vs $100 to $1000 and all the manufacturing pollution to be stuck with a GPU you don't use
link
npx 2437 days ago
I wanted to use the NVidia GPU in this laptop, it just doesn't work as well as the Intel GPU (I hate screen tearing). I think this will be great as a development environment once I get around to making Docker work with the NVidia drivers (allegedly it should). I'm not sure if you can do the same thing with AMD gear but the ability to create a Tensorflow NN and distribute it as a hardware accelerated Docker image is pretty cool.
link