• 0 Posts
  • 26 Comments
Joined 1 year ago
cake
Cake day: June 16th, 2023

help-circle



  • Ever time i see a post like this i ask the same thing and i have yet to receive answer.

    Why should i care?

    There are so many open source language models, all with different strengths and weaknesses. There are tools to run them on any OS with all kinds of different hardware requirements.

    This has been the case since before chatgpt came out and has exponentially blown up since.

    Gpt4all is just a single recent model. But in recent weeks it always gets the headlight under “run chatgpt at home”

    What does it do to stand out? Why would i use this and not one of the vicuna or llama models?

    Hugging face has a leaderboard for open source large language models.

    https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard

    If you are interested in running this tech at home, familiarize yourself with multiple models because they all will behave differently depending on your hardware and your needs.




  • What if they stored the literal files under the installation folder but then linked them under other folders classed as either Saves/configs/creations.

    In theory we just need to standardize those few classifications and people could decide where that type of files can be found, always under a folder of the name of the program.

    I feel with modern computing it doesn’t really make sense to only find certain files in a single location. The structure you use to create and save digital art may be wholy different from the one you use to browse/show/upload your art. I like to have all my game installations easily accessible for modding but i hate if i had to to use it to launch a game trough the exe. In a way other software may already cater to this need but its often bloated and far from standardized.

    While were at it allow people to set a display name for program what they like during installations and if i want to install sm multiple times, why not allow it. the os just remembers what there actually called if other software looks for it, in case of multiple it could ask which one to use.

    Of course i can dream what i want, creating new standards is probably one of the hardest thing to agree on and get done.



  • Not sure but it was a desktop version with a gui. This was on my dedicated server so not my main machine that i switched to arch. I’ve actually went to completely remove that ubuntu which was a mess from my own misdoings and experiments and started from scratch with the last LTS version of linux-server, fully in commandline. In less then a weekend i restored all the initial functionality, fixed the previously broken functionality and added some extra features to it. But again the moment i dont know how to do something i am skipping google straight for the AI genie.


  • Doing a quick google to find info about Platform Security Processor states that if you cant find the security processor section on the device security screen it means tpm is disabled. This does lead me to believe that disabling TPM at least disables windows acces to the security processor, windows cant directly use features i have directly disabled in the bios at least not without that acces.

    Or how far does this rabbit hole go exactly? I cant trow every windows device out il have to change job and my wife be pissed.



  • Oh I absolutely understand there are proper usecases for TPM like all our work laptops have bit locker enabled. But my personal device is a Diy desktop of Theseus that doesnt leave my house and it doesn’t really have all that much sensitive data anyway. My main issue with tracking/identifiers/telemetry is they use it to serve ads tailored to my behaviors they learned from the data they verified from me using those same identifiers. I am something of an anti-advertisement extremist for psychological reasons. There designed to get in my head and physically hurt.






  • No no, you read my comment wrong. I used to complain about the cli and lack of gui while trying ubuntu… with a gui.

    I am loving my arch setup. And i aint changing soon. Even if really its gpt-4 being a massive mvp to tell me how to do stuff.

    Its wasnt as much the cli stuff or any of the advanced stuff i wanted that was the problem but just that my autistic ass needed some easy/good accessible help to learn it in a way schools,google and youtube never could. Commandline is fun now and i look forward to seeing the random pokemon i get every time.



  • Hyperland is a windows manager yes because i have cognitive challenges that require visual sorting of information.

    What i gain? Super sayan levels of fast. Productivity goes brrr. Completely customizable (really into that) and it looks and feels sweet AF. This is with the hyperdot configuration found here, check out the vid. https://github.com/prasanthrangan/hyprdots

    GPT-4: it knows linux much much better then i do. I have no api so i cant just give the command box but stuff like: “provide easy to follow instructions and commands to set up x, y, z” wielded me way better result then trying the same stuff alone in linux before. I completely redid a server project i worked for more then a year on in less then a weekend. I also use it as a command cheatsheet because i suck at remembering commands and the answers on google are burried Between ads.

    Photoshop: This was a worry of myself aswell, a friend send me this “https://github.com/Gictorbit/photoshopCClinux“ Havent tried yet but its not the only option either. As i said elsewhere you can often straight up run windows installed exes from a different drive using lutrius and proton.

    I am gonna need to checkout Magnet and Raycast. They seem very promising for my job where i can only use Windows.

    Good luck if you try it! (Maybe in a vm at first)