• 0 Posts
  • 223 Comments
Joined 1 year ago
cake
Cake day: September 9th, 2023

help-circle
  • pirat@lemmy.worldtoSelfhosted@lemmy.worldLow Cost Mini PCs
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    2 hours ago

    I’m in the same situation as you, more or less… I have three new 22TB drives that need an enclosure, preferably for JBOD (no hardware RAID needed) but I can’t figure out which ones are actually good products… I don’t mind using a random-brand product if it’s actually solid.

    I find it very difficult to figure out which ones will support my 22TB drives. And for some of them, it seems, it’s impossible to add new drives to empty slots later (because of hardware RAID, I guess?), which has made me hesitant in buying one with more slots than I have drives, in case they can’t be utilized later on anyway…

    I was looking at the QNAP TR-004 which was mentioned by someone else somewhere on Lemmy some months ago, but IIRC it would be impossible to use the fourth slot later if the drive isn’t included in the hardware RAID configuration…

    EDIT: I have also been looking into so-called “backplanes” as an alternative, since they seem to do the job and are cheaper, but I’m unsure if I’ll need a PC chassis/case/tower for that to actually work?

    If you find something good (products or relevant info), feel free to share it with me.



  • Altering the prompt will certainly give a different output, though. Ok, maybe “think about this problem for a moment” is a weird prompt; I see how it actually doesn’t make much sense.

    However, including something along the lines of “think through the problem step-by-step” in the prompt really makes a difference, in my experience. The LLM will then, to a higher degree, include sections of “reasoning”, thereby arriving at an output that’s more correct or of higher quality.

    This, to me, seems like a simple precursor to the way a model like the new o1 from OpenAI (partly) works; It “thinks” about the prompt behind the scenes, presenting only the resulting output and a hidden (by default) generated summary of the secret raw “thinking” to the user.

    Of course, it’s unnecessary - maybe even stupid - to include nonsense or smalltalk in LLM prompts (unless it has proven to actually enhance the output you want), but since (some) LLMs happen to be lazy by design, telling them what to do (like reasoning) can definitely make a great difference.



  • I’m not the person you’re asking, but I have some evidence to support the case that making it work without proprietary code is a problem. GrapheneOS, a privacy-oriented Android-based smartphone OS, write in their usage guide:

    By default, GrapheneOS has always shipped with baseline support for eSIM, where users can use any eSIMs installed previously on the device. However, in order to manage and add eSIMs, proprietary Google functionality is needed. This is fully disabled by default.

    eSIM support on GrapheneOS doesn’t require any dependency on Google Play, and never shares data to Google Play even when installed.

    Edit: The fact that they haven’t implemented eSIM functionality without using the proprietary Google stuff, indicates to me that it’s either impossible/blocked or simply too hard with practically nothing in return to have been prioritised (yet?)















  • That’s movements and genres, though.

    Stop describing your paintings using sentences like “the Mona Lisa meets Monet’s water lillies”.

    or

    Stop describing your music using sentences like “similar to Bohemian Rhapsody but with drum rhythms inspired by Drop It Like It’s Hot”.

    would be somewhat more like it.

    These sentences, funnily enough, sound close to something I would write in experimental prompts for a txt2img or txt2music AI model.